LoadModelNTest.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Import package and make test environment"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Users\\lynch\\anaconda3\\envs\\DQN_win\\lib\\site-packages\\tqdm\\auto.py:22: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
      "  from .autonotebook import tqdm as notebook_tqdm\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "State shape:  (96, 96)\n",
      "Number of actions:  5\n"
     ]
    }
   ],
   "source": [
    "import time\n",
    "import gym\n",
    "from gym import envs\n",
    "from gym import wrappers, logger\n",
    "import matplotlib.pyplot as plt\n",
    "from PIL import Image\n",
    "from tqdm import tqdm\n",
    "from pathlib import Path, PureWindowsPath\n",
    "import torch\n",
    "from agent.ddpgagent import Agent3\n",
    "from agent.ppoagent import PPO\n",
    "from agent.dqnagent import Agent\n",
    "from utility import get_unique_numbers, score_save, count_impulses, count_edgezone, score_action_save\n",
    "import math\n",
    "# Device\n",
    "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
    "\n",
    "# load_tracks_from=TRACK -- Load a specific map from a pkl file\n",
    "# Make environment\n",
    "TRACK = 'track'\n",
    "env = gym.make('CarRacing-v0', num_obstacles=100, show_info_panel=1,\n",
    "               num_tracks=1, num_lanes=1, discretize_actions='soft',\n",
    "               allow_outside=False, \n",
    "               grayscale=1, frames_per_state=1,\n",
    "               load_tracks_from = TRACK)\n",
    "observation_space = env.observation_space.shape\n",
    "action_space = env.action_space.n\n",
    "print('State shape: ', observation_space)\n",
    "print('Number of actions: ', action_space)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Parameter load and create agent\n",
    "## Create storage folder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Max training time steps: 900000\t \n",
      " State dimension: (96, 96)\t \n",
      " Action size: 5\t\n",
      "K-epocs:30\n",
      " max timesteps in one episode: 1500 \n",
      " Expected episode: 1200\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from utility import Config_DQN\n",
    "from utility import Config_DDPG\n",
    "from utility import Config_PPO\n",
    "\n",
    "Episode = 1000         # test episode\n",
    "total_test_episodes = Episode\n",
    "Time_stamp = 1500   # time stamp -- simply initial parameters, no use in the test\n",
    "\n",
    "cfg_dqn = Config_DQN(print_ep=1, freq_save_epi=10, n_episodes=10, GAMMA=0.95, SEED=32, eps_end=0.001, max_t=1200)\n",
    "setattr(cfg_dqn, 'state_size', observation_space)\n",
    "setattr(cfg_dqn, 'action_size', action_space) \n",
    "\n",
    "cfg_ppo = Config_PPO(max_training_timesteps=int(900000),max_ep_len=Time_stamp,K_epochs=30, save_model_freq=500,eps_clip=0.2,gamma=0.95)\n",
    "setattr(cfg_ppo, 'state_dim', env.observation_space.shape)\n",
    "setattr(cfg_ppo, 'action_dim', env.action_space.n)\n",
    "\n",
    "print('Max training time steps: {}\\t \\n State dimension: {}\\t \\n Action size: {}\\t'.format(cfg_ppo.max_training_timesteps,cfg_ppo.state_dim,cfg_ppo.action_dim))\n",
    "print(\"K-epocs:{}\\n max timesteps in one episode: {} \\n Expected episode: {}\\n\".format(cfg_ppo.K_epochs,cfg_ppo.max_ep_len,(cfg_ppo.max_training_timesteps // cfg_ppo.max_ep_len)*2))\n",
    "\n",
    "cfg_ddpg = Config_DDPG(Episode=int(Episode), max_ep_len=Time_stamp,random_seed=int(32),freq_save_epi=int(5))\n",
    "setattr(cfg_ddpg, 'state_size', observation_space)  # create the state size parameter\n",
    "setattr(cfg_ddpg, 'action_size', action_space)      # create the action size parameter\n",
    "\n",
    "\n",
    "# Create agent\n",
    "agent_DQN = Agent(cfg_dqn)\n",
    "agent_DDPG = Agent3(cfg_ddpg)\n",
    "agent_PPO = PPO(cfg_ppo)\n",
    "\n",
    "# create folder for the test data\n",
    "file_name = 'Train_Default_II'\n",
    "p_model = (Path()/ 'document'/'Trail_Default'/file_name)\n",
    "p_model.mkdir(parents=True, exist_ok=True)\n",
    "\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Load agent and folder"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model folders\n",
    "### Model folder\n",
    "Due to variation of environment setting, please notice the printing result of:\n",
    "```python\n",
    "print(f\"DQN folders:{DQN_folders}\\n\")\n",
    "print(f\"DDPG folders:{DDPG_folders}\\n\")\n",
    "print(f\"PPO folders:{PPO_folders}\\n\")\n",
    "```\n",
    "\n",
    "The command lines print out the walking files of trained model's directory.\n",
    "\n",
    "### config folder\n",
    "Each folder(DQN, DDPG, PPO) storages the trained model with different reward configurations.\n",
    "\n",
    "```python\n",
    "print(f\"DQN model files {list(DQN_folders[0].rglob('*.pth'))}\")\n",
    "```\n",
    "This commandline prints out the walking files under the route.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Folder list in Experiment:[WindowsPath('Model/DDPG'), WindowsPath('Model/DQN'), WindowsPath('Model/PPO')]\n",
      "\n",
      "DQN folders:[WindowsPath('Model/DQN/DQN_reward_default'), WindowsPath('Model/DQN/DQN_reward_reshape')]\n",
      "\n",
      "DDPG folders:[WindowsPath('Model/DDPG/DDPG_default'), WindowsPath('Model/DDPG/DDPG_reshape_timeout')]\n",
      "\n",
      "PPO folders:[WindowsPath('Model/PPO/PPO_reshape_timeout'), WindowsPath('Model/PPO/PPO_reward_deafult')]\n",
      "\n",
      "------------------------------------\n",
      "\n",
      "DQN model files [WindowsPath('Model/DQN/DQN_reward_default/DQN_local1000.pth'), WindowsPath('Model/DQN/DQN_reward_default/DQN_local500.pth')]\n",
      "DDPG model files [WindowsPath('Model/DDPG/DDPG_default/DDPG_actor1000.pth'), WindowsPath('Model/DDPG/DDPG_default/DDPG_actor500.pth'), WindowsPath('Model/DDPG/DDPG_default/DDPG_critic1000.pth'), WindowsPath('Model/DDPG/DDPG_default/DDPG_critic500.pth')]\n",
      "PPO model files [WindowsPath('Model/PPO/PPO_reshape_timeout/PPO_adreward_1000.pth'), WindowsPath('Model/PPO/PPO_reshape_timeout/PPO_adreward_1018.pth'), WindowsPath('Model/PPO/PPO_reshape_timeout/PPO_adreward_500.pth')]\n"
     ]
    }
   ],
   "source": [
    "from pathlib import Path\n",
    "import pandas as pd\n",
    "import glob\n",
    "import os\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import math\n",
    "# All folders\n",
    "Experiment = (Path() / 'Model')\n",
    "directory = os.walk(Experiment)\n",
    "# filter out the touching the near zoen whenever the car is about to tune angle.\n",
    "gap_threshold = 35\n",
    "folder_list = [x for x in Experiment.iterdir() if x.is_dir()]\n",
    "print(f\"Folder list in Experiment:{folder_list}\\n\")\n",
    "\n",
    "# all sub folders\n",
    "DQN_folders = [a for a in folder_list[1].iterdir() if a.is_dir()]\n",
    "DDPG_folders = [b for b in folder_list[0].iterdir() if b.is_dir()]\n",
    "PPO_folders = [c for c in folder_list[2].iterdir() if c.is_dir()]\n",
    "\n",
    "# default\n",
    "print(f\"DQN folders:{DQN_folders}\\n\")\n",
    "print(f\"DDPG folders:{DDPG_folders}\\n\")\n",
    "print(f\"PPO folders:{PPO_folders}\\n\")\n",
    "print(f\"------------------------------------\\n\")\n",
    "print(f\"DQN model files {list(DQN_folders[0].rglob('*.pth'))}\")\n",
    "print(f\"DDPG model files {list(DDPG_folders[0].rglob('*.pth'))}\")\n",
    "print(f\"PPO model files {list(PPO_folders[0].rglob('*.pth'))}\")\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load Default model\n",
    "\n",
    "Find the default folder and load the trained model.\n",
    "> Hint: modify the number inside the braket: `DQN_folders[0]` (the Win-based pc might be the `0`, but might be varied due to basic setting.)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Default\n",
    "Default_DQN_model_path = DQN_folders[0]\n",
    "Default_DDPG_model_path =DDPG_folders[0]\n",
    "Default_PPO_model_path = PPO_folders[1]\n",
    "\n",
    "device = 'cuda:0' if torch.cuda.is_available() else 'cpu' # device check for GPU avaialbility\n",
    "\n",
    "# Default Agent\n",
    "DQN_model_name = f\"DQN_local{1000}.pth\"\n",
    "agent_DQN.qnetwork_local.load_state_dict(torch.load((Default_DQN_model_path/DQN_model_name),map_location=torch.device(device)))\n",
    "\n",
    "actor_name = f'DDPG_actor{1000}.pth'\n",
    "critic_name = f'DDPG_critic{1000}.pth'\n",
    "agent_DDPG.actor_local.load_state_dict(torch.load((Default_DDPG_model_path/actor_name),map_location=torch.device(device)))\n",
    "agent_DDPG.critic_local.load_state_dict(torch.load((Default_DDPG_model_path/critic_name),map_location=torch.device(device)))\n",
    "PPOmodel_name = f\"PPO_adreward_{1374}.pth\"\n",
    "agent_PPO.load((Default_PPO_model_path / PPOmodel_name))\n",
    "\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## reward reshape agent\n",
    "The opration is similar to [previous section](#load-default-model).\n",
    "\n",
    "Command loading reward reshape part if tested default agents.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DQN folders:[WindowsPath('Model/DQN/DQN_reward_default'), WindowsPath('Model/DQN/DQN_reward_reshape')]\n",
      "\n",
      "DDPG folders:[WindowsPath('Model/DDPG/DDPG_default'), WindowsPath('Model/DDPG/DDPG_reshape_timeout')]\n",
      "\n",
      "PPO folders:[WindowsPath('Model/PPO/PPO_reshape_timeout'), WindowsPath('Model/PPO/PPO_reward_deafult')]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# check folder\n",
    "print(f\"DQN folders:{DQN_folders}\\n\")\n",
    "print(f\"DDPG folders:{DDPG_folders}\\n\")\n",
    "print(f\"PPO folders:{PPO_folders}\\n\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "# # # Reshape\n",
    "# R_DQN_model_path = DQN_folders[1]\n",
    "# R_DDPG_model_path = DDPG_folders[2] # :[WindowsPath('Experiment/DDPG/DDPG_default'), WindowsPath('Experiment/DDPG/DDPG_reshape'), WindowsPath('Experiment/DDPG/DDPG_reshape_timeout')]\n",
    "# R_PPO_model_path = PPO_folders[1]\n",
    "\n",
    "# device = 'cuda:0' if torch.cuda.is_available() else 'cpu' # device check for GPU avaialbility\n",
    "\n",
    "# # Reshape Agent\n",
    "# DQN_model_name = f\"DQN_local{1000}.pth\"\n",
    "# agent_DQN.qnetwork_local.load_state_dict(torch.load((R_DQN_model_path/DQN_model_name),map_location=torch.device('cuda')))\n",
    "\n",
    "# actor_name = f'DDPG_actor{1000}.pth'\n",
    "# critic_name = f'DDPG_critic{1000}.pth'\n",
    "# agent_DDPG.actor_local.load_state_dict(torch.load((R_DDPG_model_path/actor_name),map_location=torch.device('cuda')))\n",
    "# agent_DDPG.critic_local.load_state_dict(torch.load((R_DDPG_model_path/critic_name),map_location=torch.device('cuda')))\n",
    "# PPOmodel_name = f\"PPO_adreward_{1018}.pth\"\n",
    "# agent_PPO.load((R_PPO_model_path / PPOmodel_name))\n",
    "\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Test Start"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Set up test episodes - How many testing episodes?"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## DDPG agent"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Total test episode"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# total_test_episodes = 3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Users\\lynch\\anaconda3\\envs\\DQN_win\\lib\\site-packages\\pyglet\\image\\codecs\\wic.py:406: UserWarning: [WinError -2147417850] Cannot change thread mode after it is set\n",
      "  warnings.warn(str(err))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Episode: 1 \t\t Reward: 10.95000000000121\n",
      "Episode: 2 \t\t Reward: -256.77499999999947\n",
      "Episode: 3 \t\t Reward: -186.9166666666664\n",
      "Episode: 4 \t\t Reward: -159.44999999999982\n",
      "Episode: 5 \t\t Reward: -92.95999999999924\n",
      "Episode: 6 \t\t Reward: -88.43333333333273\n",
      "Episode: 7 \t\t Reward: -156.37142857142808\n",
      "Episode: 8 \t\t Reward: -134.9124999999994\n",
      "Episode: 9 \t\t Reward: -126.16111111111061\n",
      "Episode: 10 \t\t Reward: -118.36499999999958\n",
      "Episode: 11 \t\t Reward: -111.75909090909055\n",
      "Episode: 12 \t\t Reward: -101.03749999999955\n",
      "Episode: 13 \t\t Reward: -98.64615384615344\n",
      "Episode: 14 \t\t Reward: -78.81428571428509\n",
      "Episode: 15 \t\t Reward: -111.0666666666661\n",
      "Episode: 16 \t\t Reward: -137.7124999999995\n",
      "Episode: 17 \t\t Reward: -128.01764705882294\n",
      "Episode: 18 \t\t Reward: -121.06111111111052\n",
      "Episode: 19 \t\t Reward: -119.16842105263102\n",
      "Episode: 20 \t\t Reward: -114.48499999999947\n",
      "Episode: 21 \t\t Reward: -109.75952380952329\n",
      "Episode: 22 \t\t Reward: -102.66136363636302\n",
      "Episode: 23 \t\t Reward: -98.49130434782548\n",
      "Episode: 24 \t\t Reward: -96.94791666666607\n",
      "Episode: 25 \t\t Reward: -96.52399999999943\n",
      "Episode: 26 \t\t Reward: -93.82307692307639\n",
      "Episode: 27 \t\t Reward: -90.41851851851797\n",
      "Episode: 28 \t\t Reward: -87.20535714285657\n",
      "Episode: 29 \t\t Reward: -85.26034482758567\n",
      "Episode: 30 \t\t Reward: -101.10499999999949\n",
      "Episode: 31 \t\t Reward: -99.95645161290273\n",
      "Episode: 32 \t\t Reward: -95.19531249999946\n",
      "Episode: 33 \t\t Reward: -93.26666666666615\n",
      "Episode: 34 \t\t Reward: -90.49705882352887\n",
      "Episode: 35 \t\t Reward: -87.60999999999945\n",
      "Episode: 36 \t\t Reward: -85.25277777777723\n",
      "Episode: 37 \t\t Reward: -82.4337837837832\n",
      "Episode: 38 \t\t Reward: -81.0697368421047\n",
      "Episode: 39 \t\t Reward: -93.18974358974305\n",
      "Episode: 40 \t\t Reward: -92.11124999999947\n",
      "Episode: 41 \t\t Reward: -91.08780487804827\n",
      "Episode: 42 \t\t Reward: -88.19404761904708\n",
      "Episode: 43 \t\t Reward: -98.71976744185996\n",
      "Episode: 44 \t\t Reward: -96.96704545454494\n",
      "Episode: 45 \t\t Reward: -107.50444444444393\n",
      "Episode: 46 \t\t Reward: -106.30108695652125\n",
      "Episode: 47 \t\t Reward: -105.00425531914847\n",
      "Episode: 48 \t\t Reward: -103.54479166666623\n",
      "Episode: 49 \t\t Reward: -112.88571428571385\n",
      "Episode: 50 \t\t Reward: -121.79899999999958\n",
      "Episode: 51 \t\t Reward: -120.00588235294077\n",
      "Episode: 52 \t\t Reward: -127.58173076923036\n",
      "Episode: 53 \t\t Reward: -126.69999999999962\n",
      "Episode: 54 \t\t Reward: -124.83240740740702\n",
      "Episode: 55 \t\t Reward: -123.93090909090871\n",
      "Episode: 56 \t\t Reward: -122.78660714285677\n",
      "Episode: 57 \t\t Reward: -121.62192982456105\n",
      "Episode: 58 \t\t Reward: -120.80517241379276\n",
      "Episode: 59 \t\t Reward: -118.75508474576237\n",
      "Episode: 60 \t\t Reward: -118.00916666666633\n",
      "Episode: 61 \t\t Reward: -116.2901639344259\n",
      "Episode: 62 \t\t Reward: -115.0314516129029\n",
      "Episode: 63 \t\t Reward: -112.74999999999966\n",
      "Episode: 64 \t\t Reward: -111.36718749999963\n",
      "Episode: 65 \t\t Reward: -110.50384615384579\n",
      "Episode: 66 \t\t Reward: -109.25833333333298\n",
      "Episode: 67 \t\t Reward: -108.03955223880563\n",
      "Episode: 68 \t\t Reward: -107.3272058823526\n",
      "Episode: 69 \t\t Reward: -106.41666666666633\n",
      "Episode: 70 \t\t Reward: -112.15714285714253\n",
      "Episode: 71 \t\t Reward: -110.995774647887\n",
      "Episode: 72 \t\t Reward: -109.40416666666634\n",
      "Episode: 73 \t\t Reward: -107.5547945205476\n",
      "Episode: 74 \t\t Reward: -106.84391891891859\n",
      "Episode: 75 \t\t Reward: -105.65066666666632\n",
      "Episode: 76 \t\t Reward: -101.94473684210489\n",
      "Episode: 77 \t\t Reward: -101.50064935064897\n",
      "Episode: 78 \t\t Reward: -101.15769230769193\n",
      "Episode: 79 \t\t Reward: -100.41075949367053\n",
      "Episode: 80 \t\t Reward: -106.17062499999966\n",
      "Episode: 81 \t\t Reward: -105.76419753086385\n",
      "Episode: 82 \t\t Reward: -104.96585365853626\n",
      "Episode: 83 \t\t Reward: -110.69337349397559\n",
      "Episode: 84 \t\t Reward: -110.26190476190446\n",
      "Episode: 85 \t\t Reward: -108.95999999999968\n",
      "Episode: 86 \t\t Reward: -107.69127906976713\n",
      "Episode: 87 \t\t Reward: -107.05459770114913\n",
      "Episode: 88 \t\t Reward: -105.98863636363603\n",
      "Episode: 89 \t\t Reward: -104.89494382022438\n",
      "Episode: 90 \t\t Reward: -104.02444444444413\n",
      "Episode: 91 \t\t Reward: -103.79010989010959\n",
      "Episode: 92 \t\t Reward: -103.22282608695623\n",
      "Episode: 93 \t\t Reward: -108.10215053763412\n",
      "Episode: 94 \t\t Reward: -112.42765957446778\n",
      "Episode: 95 \t\t Reward: -117.19368421052603\n",
      "Episode: 96 \t\t Reward: -115.83541666666638\n",
      "Episode: 97 \t\t Reward: -115.0737113402059\n",
      "Episode: 98 \t\t Reward: -114.05306122448951\n",
      "Episode: 99 \t\t Reward: -118.90454545454517\n",
      "Episode: 100 \t\t Reward: -117.77849999999971\n",
      "Episode: 101 \t\t Reward: -118.03399999999972\n",
      "Episode: 102 \t\t Reward: -113.5304999999997\n",
      "Episode: 103 \t\t Reward: -113.6694999999997\n",
      "Episode: 104 \t\t Reward: -113.0584999999997\n",
      "Episode: 105 \t\t Reward: -120.60349999999973\n",
      "Episode: 106 \t\t Reward: -120.29599999999974\n",
      "Episode: 107 \t\t Reward: -115.42049999999973\n",
      "Episode: 108 \t\t Reward: -116.27199999999975\n",
      "Episode: 109 \t\t Reward: -121.32599999999977\n",
      "Episode: 110 \t\t Reward: -120.64949999999975\n",
      "Episode: 111 \t\t Reward: -120.39449999999974\n",
      "Episode: 112 \t\t Reward: -121.27349999999971\n",
      "Episode: 113 \t\t Reward: -121.35049999999974\n",
      "Episode: 114 \t\t Reward: -123.70899999999976\n",
      "Episode: 115 \t\t Reward: -118.37599999999976\n",
      "Episode: 116 \t\t Reward: -113.27849999999975\n",
      "Episode: 117 \t\t Reward: -114.17799999999977\n",
      "Episode: 118 \t\t Reward: -114.36149999999978\n",
      "Episode: 119 \t\t Reward: -113.6744999999998\n",
      "Episode: 120 \t\t Reward: -113.15599999999976\n",
      "Episode: 121 \t\t Reward: -113.79549999999978\n",
      "Episode: 122 \t\t Reward: -114.72749999999982\n",
      "Episode: 123 \t\t Reward: -114.92949999999982\n",
      "Episode: 124 \t\t Reward: -114.09099999999982\n",
      "Episode: 125 \t\t Reward: -113.55049999999981\n",
      "Episode: 126 \t\t Reward: -113.44249999999978\n",
      "Episode: 127 \t\t Reward: -114.1624999999998\n",
      "Episode: 128 \t\t Reward: -119.7209999999998\n",
      "Episode: 129 \t\t Reward: -120.20699999999981\n",
      "Episode: 130 \t\t Reward: -114.5994999999998\n",
      "Episode: 131 \t\t Reward: -113.4759999999998\n",
      "Episode: 132 \t\t Reward: -114.57999999999981\n",
      "Episode: 133 \t\t Reward: -119.95299999999983\n",
      "Episode: 134 \t\t Reward: -120.46399999999983\n",
      "Episode: 135 \t\t Reward: -121.15349999999984\n",
      "Episode: 136 \t\t Reward: -121.40049999999985\n",
      "Episode: 137 \t\t Reward: -127.36249999999987\n",
      "Episode: 138 \t\t Reward: -127.86249999999987\n",
      "Episode: 139 \t\t Reward: -127.98649999999985\n",
      "Episode: 140 \t\t Reward: -127.48449999999985\n",
      "Episode: 141 \t\t Reward: -127.48299999999986\n",
      "Episode: 142 \t\t Reward: -127.84099999999985\n",
      "Episode: 143 \t\t Reward: -123.09449999999987\n",
      "Episode: 144 \t\t Reward: -122.35349999999984\n",
      "Episode: 145 \t\t Reward: -117.26699999999983\n",
      "Episode: 146 \t\t Reward: -117.41449999999985\n",
      "Episode: 147 \t\t Reward: -116.98649999999981\n",
      "Episode: 148 \t\t Reward: -117.31549999999983\n",
      "Episode: 149 \t\t Reward: -111.70149999999984\n",
      "Episode: 150 \t\t Reward: -106.88049999999983\n",
      "Episode: 151 \t\t Reward: -106.79099999999984\n",
      "Episode: 152 \t\t Reward: -102.14399999999983\n",
      "Episode: 153 \t\t Reward: -101.47949999999983\n",
      "Episode: 154 \t\t Reward: -101.87799999999982\n",
      "Episode: 155 \t\t Reward: -106.95199999999983\n",
      "Episode: 156 \t\t Reward: -106.89199999999983\n",
      "Episode: 157 \t\t Reward: -106.93849999999982\n",
      "Episode: 158 \t\t Reward: -106.5484999999998\n",
      "Episode: 159 \t\t Reward: -106.8899999999998\n",
      "Episode: 160 \t\t Reward: -106.25099999999979\n",
      "Episode: 161 \t\t Reward: -106.50899999999982\n",
      "Episode: 162 \t\t Reward: -106.79149999999981\n",
      "Episode: 163 \t\t Reward: -107.26749999999984\n",
      "Episode: 164 \t\t Reward: -107.02349999999984\n",
      "Episode: 165 \t\t Reward: -106.23699999999982\n",
      "Episode: 166 \t\t Reward: -106.75049999999982\n",
      "Episode: 167 \t\t Reward: -106.93499999999983\n",
      "Episode: 168 \t\t Reward: -106.03249999999981\n",
      "Episode: 169 \t\t Reward: -103.88749999999978\n",
      "Episode: 170 \t\t Reward: -99.41099999999977\n",
      "Episode: 171 \t\t Reward: -99.56599999999979\n",
      "Episode: 172 \t\t Reward: -99.89199999999978\n",
      "Episode: 173 \t\t Reward: -99.90799999999979\n",
      "Episode: 174 \t\t Reward: -98.96499999999976\n",
      "Episode: 175 \t\t Reward: -104.25849999999977\n",
      "Episode: 176 \t\t Reward: -106.3599999999998\n",
      "Episode: 177 \t\t Reward: -105.56199999999978\n",
      "Episode: 178 \t\t Reward: -105.30549999999977\n",
      "Episode: 179 \t\t Reward: -105.49949999999977\n",
      "Episode: 180 \t\t Reward: -99.88599999999978\n",
      "Episode: 181 \t\t Reward: -99.78849999999979\n",
      "Episode: 182 \t\t Reward: -104.62099999999975\n",
      "Episode: 183 \t\t Reward: -99.37449999999977\n",
      "Episode: 184 \t\t Reward: -98.92649999999978\n",
      "Episode: 185 \t\t Reward: -99.3469999999998\n",
      "Episode: 186 \t\t Reward: -104.43949999999978\n",
      "Episode: 187 \t\t Reward: -104.62149999999978\n",
      "Episode: 188 \t\t Reward: -105.0089999999998\n",
      "Episode: 189 \t\t Reward: -105.1519999999998\n",
      "Episode: 190 \t\t Reward: -105.49949999999978\n",
      "Episode: 191 \t\t Reward: -105.1184999999998\n",
      "Episode: 192 \t\t Reward: -105.4174999999998\n",
      "Episode: 193 \t\t Reward: -105.8089999999998\n",
      "Episode: 194 \t\t Reward: -101.13949999999981\n",
      "Episode: 195 \t\t Reward: -95.6134999999998\n",
      "Episode: 196 \t\t Reward: -96.23899999999982\n",
      "Episode: 197 \t\t Reward: -96.61499999999982\n",
      "Episode: 198 \t\t Reward: -101.68499999999982\n",
      "Episode: 199 \t\t Reward: -101.09049999999979\n",
      "Episode: 200 \t\t Reward: -101.11699999999979\n",
      "Episode: 201 \t\t Reward: -101.5984999999998\n",
      "Episode: 202 \t\t Reward: -106.6734999999998\n",
      "Episode: 203 \t\t Reward: -106.3759999999998\n",
      "Episode: 204 \t\t Reward: -106.14299999999982\n",
      "Episode: 205 \t\t Reward: -105.71249999999982\n",
      "Episode: 206 \t\t Reward: -105.35399999999981\n",
      "Episode: 207 \t\t Reward: -104.57549999999979\n",
      "Episode: 208 \t\t Reward: -109.8589999999998\n",
      "Episode: 209 \t\t Reward: -110.06599999999979\n",
      "Episode: 210 \t\t Reward: -110.76649999999981\n",
      "Episode: 211 \t\t Reward: -110.95399999999981\n",
      "Episode: 212 \t\t Reward: -108.56399999999978\n",
      "Episode: 213 \t\t Reward: -107.82249999999978\n",
      "Episode: 214 \t\t Reward: -108.07549999999979\n",
      "Episode: 215 \t\t Reward: -108.39199999999977\n",
      "Episode: 216 \t\t Reward: -108.6784999999998\n",
      "Episode: 217 \t\t Reward: -108.81599999999979\n",
      "Episode: 218 \t\t Reward: -108.83599999999979\n",
      "Episode: 219 \t\t Reward: -108.85849999999978\n",
      "Episode: 220 \t\t Reward: -114.89199999999978\n",
      "Episode: 221 \t\t Reward: -114.28449999999977\n",
      "Episode: 222 \t\t Reward: -113.97449999999976\n",
      "Episode: 223 \t\t Reward: -114.07999999999976\n",
      "Episode: 224 \t\t Reward: -114.3024999999998\n",
      "Episode: 225 \t\t Reward: -114.55749999999982\n",
      "Episode: 226 \t\t Reward: -114.67849999999981\n",
      "Episode: 227 \t\t Reward: -119.59649999999984\n",
      "Episode: 228 \t\t Reward: -114.03199999999984\n",
      "Episode: 229 \t\t Reward: -113.78899999999983\n",
      "Episode: 230 \t\t Reward: -114.46899999999984\n",
      "Episode: 231 \t\t Reward: -120.70699999999984\n",
      "Episode: 232 \t\t Reward: -120.59749999999985\n",
      "Episode: 233 \t\t Reward: -114.77199999999985\n",
      "Episode: 234 \t\t Reward: -114.75199999999984\n",
      "Episode: 235 \t\t Reward: -114.72449999999985\n",
      "Episode: 236 \t\t Reward: -113.99049999999981\n",
      "Episode: 237 \t\t Reward: -106.56899999999976\n",
      "Episode: 238 \t\t Reward: -106.02199999999974\n",
      "Episode: 239 \t\t Reward: -100.79249999999976\n",
      "Episode: 240 \t\t Reward: -100.59699999999975\n",
      "Episode: 241 \t\t Reward: -99.95149999999975\n",
      "Episode: 242 \t\t Reward: -100.41249999999977\n",
      "Episode: 243 \t\t Reward: -104.21849999999974\n",
      "Episode: 244 \t\t Reward: -105.24399999999977\n",
      "Episode: 245 \t\t Reward: -104.76199999999976\n",
      "Episode: 246 \t\t Reward: -104.85349999999976\n",
      "Episode: 247 \t\t Reward: -105.45499999999976\n",
      "Episode: 248 \t\t Reward: -105.28899999999976\n",
      "Episode: 249 \t\t Reward: -105.64099999999976\n",
      "Episode: 250 \t\t Reward: -104.86999999999976\n",
      "Episode: 251 \t\t Reward: -109.73199999999976\n",
      "Episode: 252 \t\t Reward: -109.81849999999974\n",
      "Episode: 253 \t\t Reward: -110.17699999999975\n",
      "Episode: 254 \t\t Reward: -109.71449999999976\n",
      "Episode: 255 \t\t Reward: -109.68799999999973\n",
      "Episode: 256 \t\t Reward: -109.20599999999973\n",
      "Episode: 257 \t\t Reward: -109.23849999999973\n",
      "Episode: 258 \t\t Reward: -109.77199999999975\n",
      "Episode: 259 \t\t Reward: -109.01349999999974\n",
      "Episode: 260 \t\t Reward: -114.18549999999973\n",
      "Episode: 261 \t\t Reward: -119.03599999999973\n",
      "Episode: 262 \t\t Reward: -124.04149999999973\n",
      "Episode: 263 \t\t Reward: -124.6049999999997\n",
      "Episode: 264 \t\t Reward: -125.12249999999973\n",
      "Episode: 265 \t\t Reward: -125.68899999999974\n",
      "Episode: 266 \t\t Reward: -124.48049999999974\n",
      "Episode: 267 \t\t Reward: -129.8154999999997\n",
      "Episode: 268 \t\t Reward: -129.4764999999997\n",
      "Episode: 269 \t\t Reward: -136.64999999999975\n",
      "Episode: 270 \t\t Reward: -136.06599999999972\n",
      "Episode: 271 \t\t Reward: -135.92449999999974\n",
      "Episode: 272 \t\t Reward: -141.49649999999974\n",
      "Episode: 273 \t\t Reward: -141.90949999999975\n",
      "Episode: 274 \t\t Reward: -142.3014999999998\n",
      "Episode: 275 \t\t Reward: -137.57899999999978\n",
      "Episode: 276 \t\t Reward: -137.89149999999978\n",
      "Episode: 277 \t\t Reward: -138.4389999999998\n",
      "Episode: 278 \t\t Reward: -137.9464999999998\n",
      "Episode: 279 \t\t Reward: -137.63999999999976\n",
      "Episode: 280 \t\t Reward: -137.60649999999978\n",
      "Episode: 281 \t\t Reward: -142.80499999999975\n",
      "Episode: 282 \t\t Reward: -138.13049999999978\n",
      "Episode: 283 \t\t Reward: -138.40149999999977\n",
      "Episode: 284 \t\t Reward: -143.31899999999976\n",
      "Episode: 285 \t\t Reward: -148.44999999999973\n",
      "Episode: 286 \t\t Reward: -143.85999999999976\n",
      "Episode: 287 \t\t Reward: -143.94749999999974\n",
      "Episode: 288 \t\t Reward: -144.13349999999977\n",
      "Episode: 289 \t\t Reward: -143.53949999999975\n",
      "Episode: 290 \t\t Reward: -141.2864999999997\n",
      "Episode: 291 \t\t Reward: -141.1109999999997\n",
      "Episode: 292 \t\t Reward: -140.01949999999968\n",
      "Episode: 293 \t\t Reward: -139.32699999999969\n",
      "Episode: 294 \t\t Reward: -139.4134999999997\n",
      "Episode: 295 \t\t Reward: -139.6329999999997\n",
      "Episode: 296 \t\t Reward: -138.6644999999997\n",
      "Episode: 297 \t\t Reward: -138.4769999999997\n",
      "Episode: 298 \t\t Reward: -133.49599999999973\n",
      "Episode: 299 \t\t Reward: -127.68049999999968\n",
      "Episode: 300 \t\t Reward: -128.12099999999967\n",
      "Episode: 301 \t\t Reward: -128.01649999999967\n",
      "Episode: 302 \t\t Reward: -122.83199999999967\n",
      "Episode: 303 \t\t Reward: -120.82849999999962\n",
      "Episode: 304 \t\t Reward: -121.21249999999962\n",
      "Episode: 305 \t\t Reward: -116.00999999999962\n",
      "Episode: 306 \t\t Reward: -116.39299999999962\n",
      "Episode: 307 \t\t Reward: -116.84699999999962\n",
      "Episode: 308 \t\t Reward: -110.86349999999965\n",
      "Episode: 309 \t\t Reward: -105.59299999999965\n",
      "Episode: 310 \t\t Reward: -105.36099999999966\n",
      "Episode: 311 \t\t Reward: -110.47999999999965\n",
      "Episode: 312 \t\t Reward: -112.94299999999969\n",
      "Episode: 313 \t\t Reward: -113.80199999999967\n",
      "Episode: 314 \t\t Reward: -113.71699999999969\n",
      "Episode: 315 \t\t Reward: -113.22249999999967\n",
      "Episode: 316 \t\t Reward: -112.64249999999967\n",
      "Episode: 317 \t\t Reward: -117.58249999999967\n",
      "Episode: 318 \t\t Reward: -123.32249999999968\n",
      "Episode: 319 \t\t Reward: -128.20649999999966\n",
      "Episode: 320 \t\t Reward: -122.90599999999968\n",
      "Episode: 321 \t\t Reward: -128.56199999999967\n",
      "Episode: 322 \t\t Reward: -129.2209999999997\n",
      "Episode: 323 \t\t Reward: -129.38599999999968\n",
      "Episode: 324 \t\t Reward: -135.2904999999997\n",
      "Episode: 325 \t\t Reward: -140.44149999999968\n",
      "Episode: 326 \t\t Reward: -141.07349999999968\n",
      "Episode: 327 \t\t Reward: -135.97199999999967\n",
      "Episode: 328 \t\t Reward: -136.63849999999968\n",
      "Episode: 329 \t\t Reward: -136.85949999999968\n",
      "Episode: 330 \t\t Reward: -136.5979999999997\n",
      "Episode: 331 \t\t Reward: -131.22749999999968\n",
      "Episode: 332 \t\t Reward: -131.0554999999997\n",
      "Episode: 333 \t\t Reward: -132.0399999999997\n",
      "Episode: 334 \t\t Reward: -131.50549999999967\n",
      "Episode: 335 \t\t Reward: -130.94749999999968\n",
      "Episode: 336 \t\t Reward: -131.8949999999997\n",
      "Episode: 337 \t\t Reward: -134.44649999999973\n",
      "Episode: 338 \t\t Reward: -134.73899999999972\n",
      "Episode: 339 \t\t Reward: -135.17749999999972\n",
      "Episode: 340 \t\t Reward: -141.0554999999997\n",
      "Episode: 341 \t\t Reward: -141.67499999999976\n",
      "Episode: 342 \t\t Reward: -141.30249999999972\n",
      "Episode: 343 \t\t Reward: -136.87299999999976\n",
      "Episode: 344 \t\t Reward: -136.16749999999973\n",
      "Episode: 345 \t\t Reward: -136.64049999999975\n",
      "Episode: 346 \t\t Reward: -135.68349999999975\n",
      "Episode: 347 \t\t Reward: -135.05499999999975\n",
      "Episode: 348 \t\t Reward: -134.28799999999976\n",
      "Episode: 349 \t\t Reward: -134.46299999999977\n",
      "Episode: 350 \t\t Reward: -134.77099999999976\n",
      "Episode: 351 \t\t Reward: -129.99299999999977\n",
      "Episode: 352 \t\t Reward: -130.04999999999976\n",
      "Episode: 353 \t\t Reward: -130.29349999999974\n",
      "Episode: 354 \t\t Reward: -135.90799999999973\n",
      "Episode: 355 \t\t Reward: -130.69049999999976\n",
      "Episode: 356 \t\t Reward: -131.12149999999977\n",
      "Episode: 357 \t\t Reward: -131.17499999999976\n",
      "Episode: 358 \t\t Reward: -130.83599999999976\n",
      "Episode: 359 \t\t Reward: -131.67449999999977\n",
      "Episode: 360 \t\t Reward: -127.17549999999977\n",
      "Episode: 361 \t\t Reward: -121.92099999999978\n",
      "Episode: 362 \t\t Reward: -116.54199999999977\n",
      "Episode: 363 \t\t Reward: -116.39949999999978\n",
      "Episode: 364 \t\t Reward: -116.67549999999977\n",
      "Episode: 365 \t\t Reward: -116.34149999999978\n",
      "Episode: 366 \t\t Reward: -116.9734999999998\n",
      "Episode: 367 \t\t Reward: -111.50199999999981\n",
      "Episode: 368 \t\t Reward: -112.17549999999983\n",
      "Episode: 369 \t\t Reward: -107.31799999999981\n",
      "Episode: 370 \t\t Reward: -107.29449999999983\n",
      "Episode: 371 \t\t Reward: -107.35749999999983\n",
      "Episode: 372 \t\t Reward: -107.29399999999983\n",
      "Episode: 373 \t\t Reward: -105.37099999999981\n",
      "Episode: 374 \t\t Reward: -105.5509999999998\n",
      "Episode: 375 \t\t Reward: -105.4349999999998\n",
      "Episode: 376 \t\t Reward: -105.2149999999998\n",
      "Episode: 377 \t\t Reward: -105.1309999999998\n",
      "Episode: 378 \t\t Reward: -105.5844999999998\n",
      "Episode: 379 \t\t Reward: -104.75499999999978\n",
      "Episode: 380 \t\t Reward: -105.4139999999998\n",
      "Episode: 381 \t\t Reward: -98.97199999999975\n",
      "Episode: 382 \t\t Reward: -104.23299999999976\n",
      "Episode: 383 \t\t Reward: -103.84899999999976\n",
      "Episode: 384 \t\t Reward: -98.98299999999978\n",
      "Episode: 385 \t\t Reward: -94.09999999999977\n",
      "Episode: 386 \t\t Reward: -99.37599999999976\n",
      "Episode: 387 \t\t Reward: -99.46849999999978\n",
      "Episode: 388 \t\t Reward: -98.60849999999976\n",
      "Episode: 389 \t\t Reward: -99.51799999999977\n",
      "Episode: 390 \t\t Reward: -101.8899999999998\n",
      "Episode: 391 \t\t Reward: -101.92399999999981\n",
      "Episode: 392 \t\t Reward: -102.19899999999983\n",
      "Episode: 393 \t\t Reward: -97.72349999999985\n",
      "Episode: 394 \t\t Reward: -97.76349999999982\n",
      "Episode: 395 \t\t Reward: -97.79849999999982\n",
      "Episode: 396 \t\t Reward: -99.05049999999987\n",
      "Episode: 397 \t\t Reward: -98.56499999999984\n",
      "Episode: 398 \t\t Reward: -103.33049999999984\n",
      "Episode: 399 \t\t Reward: -104.30599999999986\n",
      "Episode: 400 \t\t Reward: -104.16649999999987\n",
      "Episode: 401 \t\t Reward: -103.84249999999986\n",
      "Episode: 402 \t\t Reward: -109.06599999999985\n",
      "Episode: 403 \t\t Reward: -111.58249999999987\n",
      "Episode: 404 \t\t Reward: -112.16749999999985\n",
      "Episode: 405 \t\t Reward: -117.76999999999987\n",
      "Episode: 406 \t\t Reward: -117.92199999999988\n",
      "Episode: 407 \t\t Reward: -123.09949999999988\n",
      "Episode: 408 \t\t Reward: -123.45299999999989\n",
      "Episode: 409 \t\t Reward: -121.12099999999985\n",
      "Episode: 410 \t\t Reward: -121.76499999999983\n",
      "Episode: 411 \t\t Reward: -116.39299999999984\n",
      "Episode: 412 \t\t Reward: -115.26949999999982\n",
      "Episode: 413 \t\t Reward: -114.78749999999984\n",
      "Episode: 414 \t\t Reward: -114.17249999999984\n",
      "Episode: 415 \t\t Reward: -119.61999999999985\n",
      "Episode: 416 \t\t Reward: -119.91699999999985\n",
      "Episode: 417 \t\t Reward: -114.51299999999985\n",
      "Episode: 418 \t\t Reward: -114.47749999999985\n",
      "Episode: 419 \t\t Reward: -109.40549999999986\n",
      "Episode: 420 \t\t Reward: -108.84199999999984\n",
      "Episode: 421 \t\t Reward: -103.21649999999983\n",
      "Episode: 422 \t\t Reward: -102.56099999999984\n",
      "Episode: 423 \t\t Reward: -102.51399999999984\n",
      "Episode: 424 \t\t Reward: -97.34899999999983\n",
      "Episode: 425 \t\t Reward: -96.83599999999983\n",
      "Episode: 426 \t\t Reward: -96.66199999999982\n",
      "Episode: 427 \t\t Reward: -96.5959999999998\n",
      "Episode: 428 \t\t Reward: -95.81299999999982\n",
      "Episode: 429 \t\t Reward: -95.5784999999998\n",
      "Episode: 430 \t\t Reward: -94.56749999999978\n",
      "Episode: 431 \t\t Reward: -94.44849999999977\n",
      "Episode: 432 \t\t Reward: -94.47549999999977\n",
      "Episode: 433 \t\t Reward: -94.30299999999977\n",
      "Episode: 434 \t\t Reward: -95.19349999999979\n",
      "Episode: 435 \t\t Reward: -95.9024999999998\n",
      "Episode: 436 \t\t Reward: -94.75399999999976\n",
      "Episode: 437 \t\t Reward: -94.21249999999975\n",
      "Episode: 438 \t\t Reward: -93.83499999999977\n",
      "Episode: 439 \t\t Reward: -93.22799999999974\n",
      "Episode: 440 \t\t Reward: -87.64949999999975\n",
      "Episode: 441 \t\t Reward: -92.76249999999975\n",
      "Episode: 442 \t\t Reward: -92.87949999999975\n",
      "Episode: 443 \t\t Reward: -92.83999999999976\n",
      "Episode: 444 \t\t Reward: -92.74849999999977\n",
      "Episode: 445 \t\t Reward: -92.46399999999974\n",
      "Episode: 446 \t\t Reward: -92.46599999999977\n",
      "Episode: 447 \t\t Reward: -98.13799999999978\n",
      "Episode: 448 \t\t Reward: -98.34599999999976\n",
      "Episode: 449 \t\t Reward: -98.33199999999977\n",
      "Episode: 450 \t\t Reward: -98.91499999999976\n",
      "Episode: 451 \t\t Reward: -99.50599999999977\n",
      "Episode: 452 \t\t Reward: -99.33049999999979\n",
      "Episode: 453 \t\t Reward: -99.37449999999977\n",
      "Episode: 454 \t\t Reward: -93.29749999999976\n",
      "Episode: 455 \t\t Reward: -93.33699999999975\n",
      "Episode: 456 \t\t Reward: -93.34399999999978\n",
      "Episode: 457 \t\t Reward: -93.39599999999977\n",
      "Episode: 458 \t\t Reward: -93.53049999999978\n",
      "Episode: 459 \t\t Reward: -93.89049999999973\n",
      "Episode: 460 \t\t Reward: -98.74099999999974\n",
      "Episode: 461 \t\t Reward: -99.52799999999975\n",
      "Episode: 462 \t\t Reward: -99.55649999999976\n",
      "Episode: 463 \t\t Reward: -98.95399999999977\n",
      "Episode: 464 \t\t Reward: -98.19799999999974\n",
      "Episode: 465 \t\t Reward: -97.70299999999972\n",
      "Episode: 466 \t\t Reward: -97.2064999999997\n",
      "Episode: 467 \t\t Reward: -95.21249999999965\n",
      "Episode: 468 \t\t Reward: -100.81449999999967\n",
      "Episode: 469 \t\t Reward: -105.98449999999968\n",
      "Episode: 470 \t\t Reward: -111.35249999999968\n",
      "Episode: 471 \t\t Reward: -111.54049999999967\n",
      "Episode: 472 \t\t Reward: -106.15299999999966\n",
      "Episode: 473 \t\t Reward: -113.1349999999997\n",
      "Episode: 474 \t\t Reward: -113.72599999999973\n",
      "Episode: 475 \t\t Reward: -113.53649999999972\n",
      "Episode: 476 \t\t Reward: -113.30349999999973\n",
      "Episode: 477 \t\t Reward: -113.66149999999972\n",
      "Episode: 478 \t\t Reward: -113.57549999999972\n",
      "Episode: 479 \t\t Reward: -114.70399999999977\n",
      "Episode: 480 \t\t Reward: -114.64349999999976\n",
      "Episode: 481 \t\t Reward: -115.70449999999978\n",
      "Episode: 482 \t\t Reward: -109.61699999999978\n",
      "Episode: 483 \t\t Reward: -109.26849999999976\n",
      "Episode: 484 \t\t Reward: -109.40199999999976\n",
      "Episode: 485 \t\t Reward: -108.82949999999974\n",
      "Episode: 486 \t\t Reward: -103.74049999999974\n",
      "Episode: 487 \t\t Reward: -103.31249999999974\n",
      "Episode: 488 \t\t Reward: -109.30799999999974\n",
      "Episode: 489 \t\t Reward: -109.13299999999975\n",
      "Episode: 490 \t\t Reward: -109.15849999999975\n",
      "Episode: 491 \t\t Reward: -109.37049999999975\n",
      "Episode: 492 \t\t Reward: -109.12949999999972\n",
      "Episode: 493 \t\t Reward: -108.88349999999973\n",
      "Episode: 494 \t\t Reward: -108.74599999999971\n",
      "Episode: 495 \t\t Reward: -108.19849999999973\n",
      "Episode: 496 \t\t Reward: -108.13299999999968\n",
      "Episode: 497 \t\t Reward: -108.84149999999973\n",
      "Episode: 498 \t\t Reward: -104.01149999999973\n",
      "Episode: 499 \t\t Reward: -104.39299999999972\n",
      "Episode: 500 \t\t Reward: -104.66199999999974\n",
      "Episode: 501 \t\t Reward: -105.32099999999971\n",
      "Episode: 502 \t\t Reward: -99.48499999999972\n",
      "Episode: 503 \t\t Reward: -103.80099999999972\n",
      "Episode: 504 \t\t Reward: -103.30449999999969\n",
      "Episode: 505 \t\t Reward: -103.25699999999972\n",
      "Episode: 506 \t\t Reward: -103.01249999999973\n",
      "Episode: 507 \t\t Reward: -98.03249999999973\n",
      "Episode: 508 \t\t Reward: -98.15049999999972\n",
      "Episode: 509 \t\t Reward: -98.2704999999997\n",
      "Episode: 510 \t\t Reward: -97.3354999999997\n",
      "Episode: 511 \t\t Reward: -95.49899999999968\n",
      "Episode: 512 \t\t Reward: -96.5594999999997\n",
      "Episode: 513 \t\t Reward: -101.45699999999971\n",
      "Episode: 514 \t\t Reward: -106.9589999999997\n",
      "Episode: 515 \t\t Reward: -101.78349999999972\n",
      "Episode: 516 \t\t Reward: -101.87549999999972\n",
      "Episode: 517 \t\t Reward: -102.23649999999971\n",
      "Episode: 518 \t\t Reward: -101.62499999999972\n",
      "Episode: 519 \t\t Reward: -102.27399999999972\n",
      "Episode: 520 \t\t Reward: -102.72199999999971\n",
      "Episode: 521 \t\t Reward: -108.03049999999973\n",
      "Episode: 522 \t\t Reward: -108.19999999999973\n",
      "Episode: 523 \t\t Reward: -108.13949999999973\n",
      "Episode: 524 \t\t Reward: -107.67349999999975\n",
      "Episode: 525 \t\t Reward: -102.68699999999976\n",
      "Episode: 526 \t\t Reward: -107.02949999999973\n",
      "Episode: 527 \t\t Reward: -107.15999999999974\n",
      "Episode: 528 \t\t Reward: -107.53249999999976\n",
      "Episode: 529 \t\t Reward: -107.26699999999975\n",
      "Episode: 530 \t\t Reward: -108.72099999999976\n",
      "Episode: 531 \t\t Reward: -109.06199999999978\n",
      "Episode: 532 \t\t Reward: -108.73499999999979\n",
      "Episode: 533 \t\t Reward: -108.24249999999977\n",
      "Episode: 534 \t\t Reward: -107.79499999999979\n",
      "Episode: 535 \t\t Reward: -107.85949999999978\n",
      "Episode: 536 \t\t Reward: -108.9234999999998\n",
      "Episode: 537 \t\t Reward: -109.0604999999998\n",
      "Episode: 538 \t\t Reward: -108.8839999999998\n",
      "Episode: 539 \t\t Reward: -108.64299999999977\n",
      "Episode: 540 \t\t Reward: -113.83049999999979\n",
      "Episode: 541 \t\t Reward: -108.4419999999998\n",
      "Episode: 542 \t\t Reward: -108.78849999999979\n",
      "Episode: 543 \t\t Reward: -108.60499999999976\n",
      "Episode: 544 \t\t Reward: -109.00699999999978\n",
      "Episode: 545 \t\t Reward: -109.13999999999979\n",
      "Episode: 546 \t\t Reward: -114.38149999999979\n",
      "Episode: 547 \t\t Reward: -109.0729999999998\n",
      "Episode: 548 \t\t Reward: -109.08399999999982\n",
      "Episode: 549 \t\t Reward: -114.34299999999979\n",
      "Episode: 550 \t\t Reward: -113.9169999999998\n",
      "Episode: 551 \t\t Reward: -113.6029999999998\n",
      "Episode: 552 \t\t Reward: -112.71449999999977\n",
      "Episode: 553 \t\t Reward: -112.14399999999978\n",
      "Episode: 554 \t\t Reward: -112.7599999999998\n",
      "Episode: 555 \t\t Reward: -112.8214999999998\n",
      "Episode: 556 \t\t Reward: -112.65149999999981\n",
      "Episode: 557 \t\t Reward: -112.22799999999981\n",
      "Episode: 558 \t\t Reward: -112.0799999999998\n",
      "Episode: 559 \t\t Reward: -111.2984999999998\n",
      "Episode: 560 \t\t Reward: -111.37349999999981\n",
      "Episode: 561 \t\t Reward: -111.0634999999998\n",
      "Episode: 562 \t\t Reward: -116.36099999999982\n",
      "Episode: 563 \t\t Reward: -116.8634999999998\n",
      "Episode: 564 \t\t Reward: -117.58799999999981\n",
      "Episode: 565 \t\t Reward: -123.41699999999985\n",
      "Episode: 566 \t\t Reward: -124.48949999999986\n",
      "Episode: 567 \t\t Reward: -125.84499999999987\n",
      "Episode: 568 \t\t Reward: -120.33049999999986\n",
      "Episode: 569 \t\t Reward: -114.24449999999986\n",
      "Episode: 570 \t\t Reward: -109.60149999999985\n",
      "Episode: 571 \t\t Reward: -109.35449999999986\n",
      "Episode: 572 \t\t Reward: -109.41949999999989\n",
      "Episode: 573 \t\t Reward: -104.62949999999988\n",
      "Episode: 574 \t\t Reward: -104.30699999999987\n",
      "Episode: 575 \t\t Reward: -104.51399999999987\n",
      "Episode: 576 \t\t Reward: -104.93549999999986\n",
      "Episode: 577 \t\t Reward: -104.49399999999989\n",
      "Episode: 578 \t\t Reward: -104.82049999999987\n",
      "Episode: 579 \t\t Reward: -110.08899999999986\n",
      "Episode: 580 \t\t Reward: -109.54249999999985\n",
      "Episode: 581 \t\t Reward: -109.82199999999986\n",
      "Episode: 582 \t\t Reward: -115.74749999999986\n",
      "Episode: 583 \t\t Reward: -116.22049999999989\n",
      "Episode: 584 \t\t Reward: -116.38649999999988\n",
      "Episode: 585 \t\t Reward: -116.30349999999987\n",
      "Episode: 586 \t\t Reward: -121.59299999999989\n",
      "Episode: 587 \t\t Reward: -121.73699999999987\n",
      "Episode: 588 \t\t Reward: -116.58699999999988\n",
      "Episode: 589 \t\t Reward: -116.44749999999985\n",
      "Episode: 590 \t\t Reward: -116.14499999999985\n",
      "Episode: 591 \t\t Reward: -116.19299999999984\n",
      "Episode: 592 \t\t Reward: -114.61549999999984\n",
      "Episode: 593 \t\t Reward: -113.65849999999982\n",
      "Episode: 594 \t\t Reward: -113.90149999999981\n",
      "Episode: 595 \t\t Reward: -114.60199999999985\n",
      "Episode: 596 \t\t Reward: -114.48899999999985\n",
      "Episode: 597 \t\t Reward: -114.03349999999986\n",
      "Episode: 598 \t\t Reward: -113.83599999999984\n",
      "Episode: 599 \t\t Reward: -112.94399999999985\n",
      "Episode: 600 \t\t Reward: -112.30249999999984\n",
      "Episode: 601 \t\t Reward: -111.97049999999983\n",
      "Episode: 602 \t\t Reward: -112.66449999999985\n",
      "Episode: 603 \t\t Reward: -107.97999999999985\n",
      "Episode: 604 \t\t Reward: -107.81649999999985\n",
      "Episode: 605 \t\t Reward: -102.86849999999984\n",
      "Episode: 606 \t\t Reward: -103.30899999999983\n",
      "Episode: 607 \t\t Reward: -108.33249999999981\n",
      "Episode: 608 \t\t Reward: -113.35199999999982\n",
      "Episode: 609 \t\t Reward: -115.66949999999986\n",
      "Episode: 610 \t\t Reward: -116.32799999999989\n",
      "Episode: 611 \t\t Reward: -123.7744999999999\n",
      "Episode: 612 \t\t Reward: -122.67049999999989\n",
      "Episode: 613 \t\t Reward: -117.63599999999988\n",
      "Episode: 614 \t\t Reward: -110.31249999999986\n",
      "Episode: 615 \t\t Reward: -110.33099999999986\n",
      "Episode: 616 \t\t Reward: -115.15049999999985\n",
      "Episode: 617 \t\t Reward: -112.84749999999981\n",
      "Episode: 618 \t\t Reward: -107.51299999999979\n",
      "Episode: 619 \t\t Reward: -107.17899999999977\n",
      "Episode: 620 \t\t Reward: -107.17649999999978\n",
      "Episode: 621 \t\t Reward: -107.2049999999998\n",
      "Episode: 622 \t\t Reward: -112.44399999999979\n",
      "Episode: 623 \t\t Reward: -117.59349999999978\n",
      "Episode: 624 \t\t Reward: -117.83799999999981\n",
      "Episode: 625 \t\t Reward: -117.42649999999979\n",
      "Episode: 626 \t\t Reward: -112.8114999999998\n",
      "Episode: 627 \t\t Reward: -112.84199999999979\n",
      "Episode: 628 \t\t Reward: -113.3324999999998\n",
      "Episode: 629 \t\t Reward: -113.7724999999998\n",
      "Episode: 630 \t\t Reward: -113.32599999999981\n",
      "Episode: 631 \t\t Reward: -113.3849999999998\n",
      "Episode: 632 \t\t Reward: -114.26899999999982\n",
      "Episode: 633 \t\t Reward: -114.64949999999982\n",
      "Episode: 634 \t\t Reward: -114.96749999999982\n",
      "Episode: 635 \t\t Reward: -114.90999999999981\n",
      "Episode: 636 \t\t Reward: -114.99299999999982\n",
      "Episode: 637 \t\t Reward: -119.9594999999998\n",
      "Episode: 638 \t\t Reward: -120.55299999999983\n",
      "Episode: 639 \t\t Reward: -121.08749999999984\n",
      "Episode: 640 \t\t Reward: -116.12599999999983\n",
      "Episode: 641 \t\t Reward: -116.42149999999981\n",
      "Episode: 642 \t\t Reward: -121.10499999999983\n",
      "Episode: 643 \t\t Reward: -127.20399999999984\n",
      "Episode: 644 \t\t Reward: -132.3639999999998\n",
      "Episode: 645 \t\t Reward: -132.69499999999982\n",
      "Episode: 646 \t\t Reward: -132.76099999999983\n",
      "Episode: 647 \t\t Reward: -132.97549999999984\n",
      "Episode: 648 \t\t Reward: -138.55299999999983\n",
      "Episode: 649 \t\t Reward: -133.29249999999985\n",
      "Episode: 650 \t\t Reward: -133.43499999999986\n",
      "Episode: 651 \t\t Reward: -138.2234999999999\n",
      "Episode: 652 \t\t Reward: -139.47249999999988\n",
      "Episode: 653 \t\t Reward: -139.0164999999999\n",
      "Episode: 654 \t\t Reward: -138.99149999999986\n",
      "Episode: 655 \t\t Reward: -138.7624999999999\n",
      "Episode: 656 \t\t Reward: -139.20799999999988\n",
      "Episode: 657 \t\t Reward: -139.46599999999987\n",
      "Episode: 658 \t\t Reward: -139.55299999999986\n",
      "Episode: 659 \t\t Reward: -139.63449999999986\n",
      "Episode: 660 \t\t Reward: -134.51599999999985\n",
      "Episode: 661 \t\t Reward: -134.05199999999985\n",
      "Episode: 662 \t\t Reward: -129.18399999999986\n",
      "Episode: 663 \t\t Reward: -129.4084999999999\n",
      "Episode: 664 \t\t Reward: -134.58999999999986\n",
      "Episode: 665 \t\t Reward: -129.57649999999987\n",
      "Episode: 666 \t\t Reward: -129.16599999999988\n",
      "Episode: 667 \t\t Reward: -135.38499999999988\n",
      "Episode: 668 \t\t Reward: -135.22499999999988\n",
      "Episode: 669 \t\t Reward: -136.1804999999999\n",
      "Episode: 670 \t\t Reward: -135.5514999999999\n",
      "Episode: 671 \t\t Reward: -135.9794999999999\n",
      "Episode: 672 \t\t Reward: -135.81899999999987\n",
      "Episode: 673 \t\t Reward: -135.84199999999987\n",
      "Episode: 674 \t\t Reward: -135.5984999999999\n",
      "Episode: 675 \t\t Reward: -135.17949999999988\n",
      "Episode: 676 \t\t Reward: -134.85099999999989\n",
      "Episode: 677 \t\t Reward: -134.94399999999987\n",
      "Episode: 678 \t\t Reward: -134.7709999999999\n",
      "Episode: 679 \t\t Reward: -129.5454999999999\n",
      "Episode: 680 \t\t Reward: -130.0429999999999\n",
      "Episode: 681 \t\t Reward: -130.0104999999999\n",
      "Episode: 682 \t\t Reward: -125.2064999999999\n",
      "Episode: 683 \t\t Reward: -130.3264999999999\n",
      "Episode: 684 \t\t Reward: -129.7114999999999\n",
      "Episode: 685 \t\t Reward: -129.59999999999985\n",
      "Episode: 686 \t\t Reward: -123.88899999999987\n",
      "Episode: 687 \t\t Reward: -123.47749999999988\n",
      "Episode: 688 \t\t Reward: -122.96449999999987\n",
      "Episode: 689 \t\t Reward: -123.44049999999987\n",
      "Episode: 690 \t\t Reward: -123.35299999999988\n",
      "Episode: 691 \t\t Reward: -123.50749999999988\n",
      "Episode: 692 \t\t Reward: -129.94499999999988\n",
      "Episode: 693 \t\t Reward: -130.3529999999999\n",
      "Episode: 694 \t\t Reward: -129.77749999999992\n",
      "Episode: 695 \t\t Reward: -134.7734999999999\n",
      "Episode: 696 \t\t Reward: -139.2824999999999\n",
      "Episode: 697 \t\t Reward: -139.59599999999992\n",
      "Episode: 698 \t\t Reward: -140.17149999999992\n",
      "Episode: 699 \t\t Reward: -140.95249999999993\n",
      "Episode: 700 \t\t Reward: -141.12799999999993\n",
      "Episode: 701 \t\t Reward: -140.98299999999995\n",
      "Episode: 702 \t\t Reward: -146.16299999999993\n",
      "Episode: 703 \t\t Reward: -146.29699999999994\n",
      "Episode: 704 \t\t Reward: -146.53299999999993\n",
      "Episode: 705 \t\t Reward: -146.21649999999994\n",
      "Episode: 706 \t\t Reward: -143.88349999999988\n",
      "Episode: 707 \t\t Reward: -138.28499999999988\n",
      "Episode: 708 \t\t Reward: -138.28049999999988\n",
      "Episode: 709 \t\t Reward: -138.00849999999988\n",
      "Episode: 710 \t\t Reward: -137.4499999999999\n",
      "Episode: 711 \t\t Reward: -137.32749999999987\n",
      "Episode: 712 \t\t Reward: -138.3829999999999\n",
      "Episode: 713 \t\t Reward: -138.47399999999988\n",
      "Episode: 714 \t\t Reward: -140.59949999999992\n",
      "Episode: 715 \t\t Reward: -140.59349999999995\n",
      "Episode: 716 \t\t Reward: -136.10149999999996\n",
      "Episode: 717 \t\t Reward: -142.77949999999998\n",
      "Episode: 718 \t\t Reward: -142.88349999999997\n",
      "Episode: 719 \t\t Reward: -148.421\n",
      "Episode: 720 \t\t Reward: -147.91899999999995\n",
      "Episode: 721 \t\t Reward: -142.21699999999996\n",
      "Episode: 722 \t\t Reward: -137.26799999999992\n",
      "Episode: 723 \t\t Reward: -130.0059999999999\n",
      "Episode: 724 \t\t Reward: -129.62549999999987\n",
      "Episode: 725 \t\t Reward: -129.9174999999999\n",
      "Episode: 726 \t\t Reward: -129.74899999999988\n",
      "Episode: 727 \t\t Reward: -129.7984999999999\n",
      "Episode: 728 \t\t Reward: -129.4804999999999\n",
      "Episode: 729 \t\t Reward: -129.43449999999993\n",
      "Episode: 730 \t\t Reward: -129.5084999999999\n",
      "Episode: 731 \t\t Reward: -128.8269999999999\n",
      "Episode: 732 \t\t Reward: -133.75699999999992\n",
      "Episode: 733 \t\t Reward: -133.6729999999999\n",
      "Episode: 734 \t\t Reward: -132.9864999999999\n",
      "Episode: 735 \t\t Reward: -132.69849999999988\n",
      "Episode: 736 \t\t Reward: -132.95949999999988\n",
      "Episode: 737 \t\t Reward: -133.25749999999988\n",
      "Episode: 738 \t\t Reward: -132.43849999999986\n",
      "Episode: 739 \t\t Reward: -132.63699999999986\n",
      "Episode: 740 \t\t Reward: -132.43449999999987\n",
      "Episode: 741 \t\t Reward: -132.74599999999987\n",
      "Episode: 742 \t\t Reward: -127.53849999999984\n",
      "Episode: 743 \t\t Reward: -121.62299999999986\n",
      "Episode: 744 \t\t Reward: -116.46199999999985\n",
      "Episode: 745 \t\t Reward: -116.10749999999986\n",
      "Episode: 746 \t\t Reward: -110.48449999999983\n",
      "Episode: 747 \t\t Reward: -110.24699999999983\n",
      "Episode: 748 \t\t Reward: -105.07449999999984\n",
      "Episode: 749 \t\t Reward: -105.42749999999982\n",
      "Episode: 750 \t\t Reward: -110.58149999999982\n",
      "Episode: 751 \t\t Reward: -110.92899999999982\n",
      "Episode: 752 \t\t Reward: -110.20599999999982\n",
      "Episode: 753 \t\t Reward: -114.94749999999982\n",
      "Episode: 754 \t\t Reward: -114.85949999999983\n",
      "Episode: 755 \t\t Reward: -114.7234999999998\n",
      "Episode: 756 \t\t Reward: -114.4979999999998\n",
      "Episode: 757 \t\t Reward: -114.52299999999981\n",
      "Episode: 758 \t\t Reward: -114.03349999999982\n",
      "Episode: 759 \t\t Reward: -114.60299999999982\n",
      "Episode: 760 \t\t Reward: -114.56149999999984\n",
      "Episode: 761 \t\t Reward: -114.95649999999983\n",
      "Episode: 762 \t\t Reward: -114.86899999999982\n",
      "Episode: 763 \t\t Reward: -114.69549999999983\n",
      "Episode: 764 \t\t Reward: -108.75199999999984\n",
      "Episode: 765 \t\t Reward: -109.17149999999981\n",
      "Episode: 766 \t\t Reward: -109.49599999999982\n",
      "Episode: 767 \t\t Reward: -104.37599999999982\n",
      "Episode: 768 \t\t Reward: -105.01149999999986\n",
      "Episode: 769 \t\t Reward: -102.58599999999979\n",
      "Episode: 770 \t\t Reward: -102.1424999999998\n",
      "Episode: 771 \t\t Reward: -102.02499999999976\n",
      "Episode: 772 \t\t Reward: -102.1214999999998\n",
      "Episode: 773 \t\t Reward: -102.30949999999977\n",
      "Episode: 774 \t\t Reward: -102.94899999999976\n",
      "Episode: 775 \t\t Reward: -102.72049999999977\n",
      "Episode: 776 \t\t Reward: -103.14399999999978\n",
      "Episode: 777 \t\t Reward: -103.25549999999977\n",
      "Episode: 778 \t\t Reward: -103.43199999999977\n",
      "Episode: 779 \t\t Reward: -102.26049999999974\n",
      "Episode: 780 \t\t Reward: -107.27399999999975\n",
      "Episode: 781 \t\t Reward: -112.51049999999975\n",
      "Episode: 782 \t\t Reward: -112.45699999999975\n",
      "Episode: 783 \t\t Reward: -107.28399999999978\n",
      "Episode: 784 \t\t Reward: -107.30749999999975\n",
      "Episode: 785 \t\t Reward: -107.80199999999977\n",
      "Episode: 786 \t\t Reward: -105.87549999999972\n",
      "Episode: 787 \t\t Reward: -106.34799999999974\n",
      "Episode: 788 \t\t Reward: -106.43649999999975\n",
      "Episode: 789 \t\t Reward: -111.43399999999974\n",
      "Episode: 790 \t\t Reward: -111.06499999999974\n",
      "Episode: 791 \t\t Reward: -110.34349999999971\n",
      "Episode: 792 \t\t Reward: -105.75849999999973\n",
      "Episode: 793 \t\t Reward: -104.06999999999971\n",
      "Episode: 794 \t\t Reward: -109.82899999999971\n",
      "Episode: 795 \t\t Reward: -104.38349999999971\n",
      "Episode: 796 \t\t Reward: -104.4359999999997\n",
      "Episode: 797 \t\t Reward: -103.82849999999969\n",
      "Episode: 798 \t\t Reward: -108.58099999999969\n",
      "Episode: 799 \t\t Reward: -108.40399999999968\n",
      "Episode: 800 \t\t Reward: -108.6819999999997\n",
      "Episode: 801 \t\t Reward: -108.06999999999965\n",
      "Episode: 802 \t\t Reward: -102.94899999999967\n",
      "Episode: 803 \t\t Reward: -102.02699999999965\n",
      "Episode: 804 \t\t Reward: -101.92899999999965\n",
      "Episode: 805 \t\t Reward: -102.26699999999964\n",
      "Episode: 806 \t\t Reward: -104.33999999999969\n",
      "Episode: 807 \t\t Reward: -109.98599999999969\n",
      "Episode: 808 \t\t Reward: -104.70849999999969\n",
      "Episode: 809 \t\t Reward: -104.73849999999972\n",
      "Episode: 810 \t\t Reward: -110.3644999999997\n",
      "Episode: 811 \t\t Reward: -105.00449999999972\n",
      "Episode: 812 \t\t Reward: -104.7464999999997\n",
      "Episode: 813 \t\t Reward: -109.4304999999997\n",
      "Episode: 814 \t\t Reward: -109.6914999999997\n",
      "Episode: 815 \t\t Reward: -109.9429999999997\n",
      "Episode: 816 \t\t Reward: -109.88899999999968\n",
      "Episode: 817 \t\t Reward: -110.28249999999969\n",
      "Episode: 818 \t\t Reward: -110.88199999999969\n",
      "Episode: 819 \t\t Reward: -110.8744999999997\n",
      "Episode: 820 \t\t Reward: -116.73649999999972\n",
      "Episode: 821 \t\t Reward: -117.71499999999975\n",
      "Episode: 822 \t\t Reward: -122.99549999999974\n",
      "Episode: 823 \t\t Reward: -124.67399999999978\n",
      "Episode: 824 \t\t Reward: -129.8524999999998\n",
      "Episode: 825 \t\t Reward: -129.7284999999998\n",
      "Episode: 826 \t\t Reward: -129.88949999999974\n",
      "Episode: 827 \t\t Reward: -129.20049999999978\n",
      "Episode: 828 \t\t Reward: -134.58999999999975\n",
      "Episode: 829 \t\t Reward: -134.69649999999976\n",
      "Episode: 830 \t\t Reward: -134.77899999999974\n",
      "Episode: 831 \t\t Reward: -135.04449999999977\n",
      "Episode: 832 \t\t Reward: -133.90699999999975\n",
      "Episode: 833 \t\t Reward: -133.92649999999975\n",
      "Episode: 834 \t\t Reward: -134.72199999999978\n",
      "Episode: 835 \t\t Reward: -132.54549999999972\n",
      "Episode: 836 \t\t Reward: -132.40849999999972\n",
      "Episode: 837 \t\t Reward: -126.4429999999997\n",
      "Episode: 838 \t\t Reward: -127.13049999999971\n",
      "Episode: 839 \t\t Reward: -126.48849999999972\n",
      "Episode: 840 \t\t Reward: -127.21949999999971\n",
      "Episode: 841 \t\t Reward: -126.57549999999972\n",
      "Episode: 842 \t\t Reward: -126.6364999999997\n",
      "Episode: 843 \t\t Reward: -126.34249999999973\n",
      "Episode: 844 \t\t Reward: -126.80449999999969\n",
      "Episode: 845 \t\t Reward: -126.81899999999972\n",
      "Episode: 846 \t\t Reward: -127.74599999999975\n",
      "Episode: 847 \t\t Reward: -127.79799999999976\n",
      "Episode: 848 \t\t Reward: -127.99549999999974\n",
      "Episode: 849 \t\t Reward: -127.58349999999975\n",
      "Episode: 850 \t\t Reward: -122.16449999999973\n",
      "Episode: 851 \t\t Reward: -122.20399999999974\n",
      "Episode: 852 \t\t Reward: -122.79649999999972\n",
      "Episode: 853 \t\t Reward: -117.83799999999974\n",
      "Episode: 854 \t\t Reward: -117.94699999999973\n",
      "Episode: 855 \t\t Reward: -118.07349999999975\n",
      "Episode: 856 \t\t Reward: -117.81699999999974\n",
      "Episode: 857 \t\t Reward: -116.73899999999972\n",
      "Episode: 858 \t\t Reward: -122.52899999999973\n",
      "Episode: 859 \t\t Reward: -122.06499999999973\n",
      "Episode: 860 \t\t Reward: -121.41949999999972\n",
      "Episode: 861 \t\t Reward: -121.0189999999997\n",
      "Episode: 862 \t\t Reward: -120.70049999999968\n",
      "Episode: 863 \t\t Reward: -120.47249999999971\n",
      "Episode: 864 \t\t Reward: -120.62049999999972\n",
      "Episode: 865 \t\t Reward: -120.28749999999971\n",
      "Episode: 866 \t\t Reward: -119.73749999999973\n",
      "Episode: 867 \t\t Reward: -118.66199999999968\n",
      "Episode: 868 \t\t Reward: -117.81149999999965\n",
      "Episode: 869 \t\t Reward: -124.80349999999967\n",
      "Episode: 870 \t\t Reward: -130.8909999999997\n",
      "Episode: 871 \t\t Reward: -130.6024999999997\n",
      "Episode: 872 \t\t Reward: -135.3254999999997\n",
      "Episode: 873 \t\t Reward: -140.4249999999997\n",
      "Episode: 874 \t\t Reward: -140.27399999999972\n",
      "Episode: 875 \t\t Reward: -141.17199999999968\n",
      "Episode: 876 \t\t Reward: -141.3139999999997\n",
      "Episode: 877 \t\t Reward: -141.00849999999969\n",
      "Episode: 878 \t\t Reward: -140.4174999999997\n",
      "Episode: 879 \t\t Reward: -140.9944999999997\n",
      "Episode: 880 \t\t Reward: -135.78249999999971\n",
      "Episode: 881 \t\t Reward: -130.2429999999997\n",
      "Episode: 882 \t\t Reward: -129.81349999999972\n",
      "Episode: 883 \t\t Reward: -129.85849999999974\n",
      "Episode: 884 \t\t Reward: -130.41549999999972\n",
      "Episode: 885 \t\t Reward: -128.2734999999997\n",
      "Episode: 886 \t\t Reward: -129.87949999999972\n",
      "Episode: 887 \t\t Reward: -129.5424999999997\n",
      "Episode: 888 \t\t Reward: -129.5379999999997\n",
      "Episode: 889 \t\t Reward: -129.60149999999973\n",
      "Episode: 890 \t\t Reward: -129.85749999999973\n",
      "Episode: 891 \t\t Reward: -130.43549999999976\n",
      "Episode: 892 \t\t Reward: -130.66799999999975\n",
      "Episode: 893 \t\t Reward: -132.8059999999998\n",
      "Episode: 894 \t\t Reward: -132.65549999999976\n",
      "Episode: 895 \t\t Reward: -132.74799999999976\n",
      "Episode: 896 \t\t Reward: -126.90949999999975\n",
      "Episode: 897 \t\t Reward: -127.16649999999976\n",
      "Episode: 898 \t\t Reward: -122.42699999999977\n",
      "Episode: 899 \t\t Reward: -122.08599999999974\n",
      "Episode: 900 \t\t Reward: -122.40249999999976\n",
      "Episode: 901 \t\t Reward: -122.97649999999977\n",
      "Episode: 902 \t\t Reward: -122.06299999999973\n",
      "Episode: 903 \t\t Reward: -122.68699999999976\n",
      "Episode: 904 \t\t Reward: -127.77449999999978\n",
      "Episode: 905 \t\t Reward: -127.39599999999977\n",
      "Episode: 906 \t\t Reward: -126.82999999999976\n",
      "Episode: 907 \t\t Reward: -121.88349999999977\n",
      "Episode: 908 \t\t Reward: -122.46499999999978\n",
      "Episode: 909 \t\t Reward: -122.5449999999998\n",
      "Episode: 910 \t\t Reward: -117.41999999999979\n",
      "Episode: 911 \t\t Reward: -117.69249999999977\n",
      "Episode: 912 \t\t Reward: -122.46449999999976\n",
      "Episode: 913 \t\t Reward: -118.03999999999975\n",
      "Episode: 914 \t\t Reward: -117.35199999999975\n",
      "Episode: 915 \t\t Reward: -117.15099999999977\n",
      "Episode: 916 \t\t Reward: -116.90449999999977\n",
      "Episode: 917 \t\t Reward: -111.38799999999975\n",
      "Episode: 918 \t\t Reward: -111.37449999999976\n",
      "Episode: 919 \t\t Reward: -106.32749999999976\n",
      "Episode: 920 \t\t Reward: -100.36449999999974\n",
      "Episode: 921 \t\t Reward: -100.05749999999975\n",
      "Episode: 922 \t\t Reward: -94.53999999999972\n",
      "Episode: 923 \t\t Reward: -94.52099999999973\n",
      "Episode: 924 \t\t Reward: -89.6789999999997\n",
      "Episode: 925 \t\t Reward: -95.53949999999972\n",
      "Episode: 926 \t\t Reward: -94.75349999999973\n",
      "Episode: 927 \t\t Reward: -95.06399999999972\n",
      "Episode: 928 \t\t Reward: -90.04799999999972\n",
      "Episode: 929 \t\t Reward: -88.76249999999969\n",
      "Episode: 930 \t\t Reward: -89.0294999999997\n",
      "Episode: 931 \t\t Reward: -89.0194999999997\n",
      "Episode: 932 \t\t Reward: -85.02549999999972\n",
      "Episode: 933 \t\t Reward: -85.01799999999972\n",
      "Episode: 934 \t\t Reward: -84.19899999999973\n",
      "Episode: 935 \t\t Reward: -86.09599999999975\n",
      "Episode: 936 \t\t Reward: -86.15099999999973\n",
      "Episode: 937 \t\t Reward: -86.35349999999976\n",
      "Episode: 938 \t\t Reward: -86.72599999999974\n",
      "Episode: 939 \t\t Reward: -86.61099999999976\n",
      "Episode: 940 \t\t Reward: -91.52899999999975\n",
      "Episode: 941 \t\t Reward: -91.48599999999976\n",
      "Episode: 942 \t\t Reward: -97.02949999999977\n",
      "Episode: 943 \t\t Reward: -97.09499999999979\n",
      "Episode: 944 \t\t Reward: -97.06149999999978\n",
      "Episode: 945 \t\t Reward: -102.23849999999976\n",
      "Episode: 946 \t\t Reward: -102.22449999999976\n",
      "Episode: 947 \t\t Reward: -102.44749999999978\n",
      "Episode: 948 \t\t Reward: -102.30099999999979\n",
      "Episode: 949 \t\t Reward: -107.49899999999975\n",
      "Episode: 950 \t\t Reward: -107.75449999999975\n",
      "Episode: 951 \t\t Reward: -102.00249999999977\n",
      "Episode: 952 \t\t Reward: -102.00999999999976\n",
      "Episode: 953 \t\t Reward: -103.11299999999977\n",
      "Episode: 954 \t\t Reward: -102.7674999999998\n",
      "Episode: 955 \t\t Reward: -103.08749999999978\n",
      "Episode: 956 \t\t Reward: -108.56149999999981\n",
      "Episode: 957 \t\t Reward: -109.65499999999982\n",
      "Episode: 958 \t\t Reward: -103.27799999999979\n",
      "Episode: 959 \t\t Reward: -103.75649999999979\n",
      "Episode: 960 \t\t Reward: -104.3984999999998\n",
      "Episode: 961 \t\t Reward: -105.2109999999998\n",
      "Episode: 962 \t\t Reward: -104.8689999999998\n",
      "Episode: 963 \t\t Reward: -110.2824999999998\n",
      "Episode: 964 \t\t Reward: -110.6884999999998\n",
      "Episode: 965 \t\t Reward: -110.5734999999998\n",
      "Episode: 966 \t\t Reward: -110.41199999999981\n",
      "Episode: 967 \t\t Reward: -111.28799999999981\n",
      "Episode: 968 \t\t Reward: -111.88699999999984\n",
      "Episode: 969 \t\t Reward: -107.29199999999986\n",
      "Episode: 970 \t\t Reward: -101.88749999999986\n",
      "Episode: 971 \t\t Reward: -102.32999999999986\n",
      "Episode: 972 \t\t Reward: -97.94499999999985\n",
      "Episode: 973 \t\t Reward: -92.19099999999986\n",
      "Episode: 974 \t\t Reward: -97.06649999999983\n",
      "Episode: 975 \t\t Reward: -96.40649999999985\n",
      "Episode: 976 \t\t Reward: -96.21499999999986\n",
      "Episode: 977 \t\t Reward: -95.84349999999984\n",
      "Episode: 978 \t\t Reward: -95.68599999999982\n",
      "Episode: 979 \t\t Reward: -95.5084999999998\n",
      "Episode: 980 \t\t Reward: -95.23699999999981\n",
      "Episode: 981 \t\t Reward: -95.15799999999979\n",
      "Episode: 982 \t\t Reward: -99.43499999999977\n",
      "Episode: 983 \t\t Reward: -98.82099999999978\n",
      "Episode: 984 \t\t Reward: -97.65849999999975\n",
      "Episode: 985 \t\t Reward: -100.01499999999979\n",
      "Episode: 986 \t\t Reward: -99.95649999999979\n",
      "Episode: 987 \t\t Reward: -104.22049999999976\n",
      "Episode: 988 \t\t Reward: -104.70499999999976\n",
      "Episode: 989 \t\t Reward: -98.73549999999976\n",
      "Episode: 990 \t\t Reward: -98.73849999999976\n",
      "Episode: 991 \t\t Reward: -98.43899999999977\n",
      "Episode: 992 \t\t Reward: -102.96999999999974\n",
      "Episode: 993 \t\t Reward: -103.01699999999974\n",
      "Episode: 994 \t\t Reward: -97.27299999999974\n",
      "Episode: 995 \t\t Reward: -97.40999999999974\n",
      "Episode: 996 \t\t Reward: -98.76799999999977\n",
      "Episode: 997 \t\t Reward: -98.84699999999977\n",
      "Episode: 998 \t\t Reward: -99.02699999999977\n",
      "Episode: 999 \t\t Reward: -104.59099999999977\n",
      "Episode: 1000 \t\t Reward: -104.04599999999975\n",
      "Total test time: 02:38:33.30\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from collections import deque\n",
    "import numpy as np\n",
    "from utility import get_unique_numbers, count_edgezone\n",
    "import time\n",
    "# DDPG\n",
    "\n",
    "test_running_reward = 0 # reset calculation reward\n",
    "\n",
    "ac =[]\n",
    "start = time.time()\n",
    "runn_ep =[]\n",
    "scores =[]\n",
    "scores_deque = deque(maxlen=100)\n",
    "avg_scores =[]\n",
    "Bmp = []\n",
    "T_Bmp = []\n",
    "bumps = 0\n",
    "if p_model.is_dir():\n",
    "    nb_name = 'DDPG_Test_log.txt'\n",
    "with open(p_model.joinpath(nb_name), \"w\") as f:\n",
    "\n",
    "    for ep in range(1, total_test_episodes+1):\n",
    "        \n",
    "        start_ep = time.time()\n",
    "        ep_reward = 0\n",
    "        runn_ep.append(ep)\n",
    "        state = env.reset()\n",
    "        Ep_bmp = []\n",
    "        for t in range(1, 1500):\n",
    "            # bumps = 0\n",
    "            action = agent_DDPG.act(state)\n",
    "            ac.append(action)\n",
    "            state, reward, done, bumps = env.step(action) # reward return\n",
    "\n",
    "            Ep_bmp.append(bumps)\n",
    "            T_Bmp.append(bumps)\n",
    "            ep_reward += reward\n",
    "            # env.render('rgb_array') # uncommand to display -- warning: expensive resources.\n",
    "\n",
    "            if done:\n",
    "                break\n",
    "        Bmp.append(Ep_bmp)\n",
    "        scores_deque.append(ep_reward)\n",
    "        avg_scores.append(np.mean(scores_deque))\n",
    "        end_ep = time.time()\n",
    "        hours_ep, rem_ep = divmod(end_ep-start_ep, 3600)\n",
    "        minutes_ep, seconds_ep = divmod(rem_ep, 60)\n",
    "        \n",
    "\n",
    "        f.write(\"Episode{} (survive time): {:0>2}:{:0>2}:{:05.2f}\\n\".format(ep,int(hours_ep),int(minutes_ep),seconds_ep))\n",
    "        test_running_reward +=  ep_reward\n",
    "\n",
    "        # print('Episode: {} \\t\\t Reward: {}'.format(ep, round(ep_reward, 2)))\n",
    "        print('Episode: {} \\t\\t Reward: {}'.format(ep, avg_scores[ep-1]))\n",
    "\n",
    "        scores.append(round(ep_reward, 2))\n",
    "        ep_reward = 0\n",
    "    env.close()\n",
    "\n",
    "    end = time.time()\n",
    "    hours, rem = divmod(end-start, 3600)\n",
    "    minutes, seconds = divmod(rem, 60)\n",
    "    print(\"Total test time: {:0>2}:{:0>2}:{:05.2f}\".format(int(hours),int(minutes),seconds))\n",
    "    \n",
    "    # count the bump instances.\n",
    "    count_list_DDPG = []\n",
    "    for i in range(len(Bmp)):\n",
    "        count_list_DDPG.append(count_impulses(Bmp[i]))\n",
    "    data_ddpg = {'Bump_instance': count_list_DDPG}\n",
    "    Bump_track = pd.DataFrame(data_ddpg)\n",
    "    Tcount_DDPG = count_edgezone(T_Bmp,gap_threshold )\n",
    "    \n",
    "    f.write(\"Total Test time (survive time): {:0>2}:{:0>2}:{:05.2f}\\n\".format(int(hours),int(minutes),seconds))\n",
    "    f.write('Toal epoch:%s\\n'%total_test_episodes)\n",
    "    f.write(\"Bump instances of total: %s\\n\"%Tcount_DDPG)\n",
    "    f.write(\"Bump to edge zone %s\\n\"%count_edgezone(T_Bmp, gap_threshold))\n",
    "    f.close()\n",
    "\n",
    "\n",
    "DDPG_avg = avg_scores\n",
    "DDPG_scores = scores\n",
    "DDPG_ac = ac\n",
    "\n",
    "# save csv file\n",
    "csv_name = f\"DDPG_default\"\n",
    "score_action_save(DDPG_ac,DDPG_scores,DDPG_avg,file_name=csv_name,path =p_model,print_=False) # save score table as csv file\n",
    "# score_save(scores,avg_scores,file_name=csv_name,path=p_model,print_=False) # save score table as csv file\n",
    "count_impulse_file = f\"DDPG_default_count\"\n",
    "Bump_track.to_csv((p_model/f\"{count_impulse_file}.csv\")) # save bump instances as csv file\n",
    "\n",
    "\n",
    "# # plot the bump instances.\n",
    "# for i in range(total_test_episodes):\n",
    "#     plt.subplot(1, total_test_episodes, i+1)\n",
    "#     plt.plot(Bmp[i])\n",
    "#     plt.title(f'Ct_:{data[\"Bump_instance\"][i]}')\n",
    "\n",
    "# plot for touch instanaces for the track.\n",
    "# fig = plt.figure()\n",
    "# plt.plot(T_Bmp)\n",
    "# plt.title(f'Total bump instances:{Tcount_DDPG}')\n",
    "\n",
    "fig = plt.figure()\n",
    "ax = fig.add_subplot(111)\n",
    "plt.plot(avg_scores)\n",
    "# plt.plot(df['Averagescores'])\n",
    "plt.ylabel('Scores')\n",
    "plt.xlabel(f'Episode {len(avg_scores)}')\n",
    "plt.show()\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "\n",
    "action_list = [0,1,2,3,4]\n",
    "h=[]\n",
    "for i in action_list:\n",
    "    h.append(ac.count(i))\n",
    "bar_contain = plt.bar(action_list,h)\n",
    "plt.xlabel('Action items')\n",
    "plt.ylabel('Number of action item used during trial')\n",
    "plt.bar_label(bar_contain)\n",
    "plt.show()\n",
    "\n",
    "\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## PPO agent test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "807038"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import gc\n",
    "\n",
    "# del agent_DQN\n",
    "del agent_DDPG\n",
    "# del agent_PPO\n",
    "torch.cuda.empty_cache()\n",
    "gc.collect()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Episode: 1 \t\t Reward: 0.15\n",
      "Episode: 2 \t\t Reward: -8.749999999999847\n",
      "Episode: 3 \t\t Reward: 2.350000000000622\n",
      "Episode: 4 \t\t Reward: -8.82499999999961\n",
      "Episode: 5 \t\t Reward: -121.70999999999972\n",
      "Episode: 6 \t\t Reward: -101.39999999999976\n",
      "Episode: 7 \t\t Reward: -159.3571428571426\n",
      "Episode: 8 \t\t Reward: -211.72499999999977\n",
      "Episode: 9 \t\t Reward: -191.3499999999998\n",
      "Episode: 10 \t\t Reward: -154.51499999999947\n",
      "Episode: 11 \t\t Reward: -145.77272727272683\n",
      "Episode: 12 \t\t Reward: -135.47499999999957\n",
      "Episode: 13 \t\t Reward: -125.89999999999957\n",
      "Episode: 14 \t\t Reward: -115.28928571428521\n",
      "Episode: 15 \t\t Reward: -111.67666666666621\n",
      "Episode: 16 \t\t Reward: -104.3999999999995\n",
      "Episode: 17 \t\t Reward: -132.11176470588188\n",
      "Episode: 18 \t\t Reward: -126.26944444444399\n",
      "Episode: 19 \t\t Reward: -116.21842105263099\n",
      "Episode: 20 \t\t Reward: -114.11749999999945\n",
      "Episode: 21 \t\t Reward: -107.84285714285654\n",
      "Episode: 22 \t\t Reward: -128.87499999999943\n",
      "Episode: 23 \t\t Reward: -120.76956521739066\n",
      "Episode: 24 \t\t Reward: -115.71874999999937\n",
      "Episode: 25 \t\t Reward: -114.4259999999994\n",
      "Episode: 26 \t\t Reward: -109.90769230769169\n",
      "Episode: 27 \t\t Reward: -108.05925925925867\n",
      "Episode: 28 \t\t Reward: -122.39285714285656\n",
      "Episode: 29 \t\t Reward: -136.91206896551668\n",
      "Episode: 30 \t\t Reward: -134.09833333333282\n",
      "Episode: 31 \t\t Reward: -130.6306451612898\n",
      "Episode: 32 \t\t Reward: -128.40624999999952\n",
      "Episode: 33 \t\t Reward: -125.18939393939347\n",
      "Episode: 34 \t\t Reward: -120.99264705882305\n",
      "Episode: 35 \t\t Reward: -132.84285714285667\n",
      "Episode: 36 \t\t Reward: -127.73333333333282\n",
      "Episode: 37 \t\t Reward: -138.35945945945895\n",
      "Episode: 38 \t\t Reward: -136.83157894736792\n",
      "Episode: 39 \t\t Reward: -134.14871794871746\n",
      "Episode: 40 \t\t Reward: -131.20499999999953\n",
      "Episode: 41 \t\t Reward: -127.02195121951168\n",
      "Episode: 42 \t\t Reward: -136.79880952380904\n",
      "Episode: 43 \t\t Reward: -146.9790697674414\n",
      "Episode: 44 \t\t Reward: -144.29886363636317\n",
      "Episode: 45 \t\t Reward: -140.1788888888884\n",
      "Episode: 46 \t\t Reward: -137.12826086956474\n",
      "Episode: 47 \t\t Reward: -143.5723404255314\n",
      "Episode: 48 \t\t Reward: -151.81354166666614\n",
      "Episode: 49 \t\t Reward: -158.57448979591783\n",
      "Episode: 50 \t\t Reward: -166.6429999999995\n",
      "Episode: 51 \t\t Reward: -159.88529411764648\n",
      "Episode: 52 \t\t Reward: -158.3971153846148\n",
      "Episode: 53 \t\t Reward: -156.1764150943391\n",
      "Episode: 54 \t\t Reward: -163.21666666666613\n",
      "Episode: 55 \t\t Reward: -169.8363636363631\n",
      "Episode: 56 \t\t Reward: -167.67053571428522\n",
      "Episode: 57 \t\t Reward: -174.63684210526267\n",
      "Episode: 58 \t\t Reward: -181.54999999999953\n",
      "Episode: 59 \t\t Reward: -179.60593220338936\n",
      "Episode: 60 \t\t Reward: -177.62499999999955\n",
      "Episode: 61 \t\t Reward: -175.10327868852414\n",
      "Episode: 62 \t\t Reward: -173.30403225806404\n",
      "Episode: 63 \t\t Reward: -170.16190476190428\n",
      "Episode: 64 \t\t Reward: -176.63593749999956\n",
      "Episode: 65 \t\t Reward: -174.87153846153802\n",
      "Episode: 66 \t\t Reward: -172.0651515151511\n",
      "Episode: 67 \t\t Reward: -169.40223880596972\n",
      "Episode: 68 \t\t Reward: -167.2985294117643\n",
      "Episode: 69 \t\t Reward: -165.94782608695613\n",
      "Episode: 70 \t\t Reward: -164.0999999999996\n",
      "Episode: 71 \t\t Reward: -162.38028169014046\n",
      "Episode: 72 \t\t Reward: -160.85555555555513\n",
      "Episode: 73 \t\t Reward: -159.21369863013658\n",
      "Episode: 74 \t\t Reward: -164.5749999999996\n",
      "Episode: 75 \t\t Reward: -162.37866666666628\n",
      "Episode: 76 \t\t Reward: -161.1230263157891\n",
      "Episode: 77 \t\t Reward: -166.54415584415545\n",
      "Episode: 78 \t\t Reward: -163.99807692307652\n",
      "Episode: 79 \t\t Reward: -169.26772151898695\n",
      "Episode: 80 \t\t Reward: -167.55437499999962\n",
      "Episode: 81 \t\t Reward: -172.14629629629593\n",
      "Episode: 82 \t\t Reward: -170.84512195121914\n",
      "Episode: 83 \t\t Reward: -169.53012048192736\n",
      "Episode: 84 \t\t Reward: -166.69345238095198\n",
      "Episode: 85 \t\t Reward: -170.35470588235253\n",
      "Episode: 86 \t\t Reward: -167.74825581395305\n",
      "Episode: 87 \t\t Reward: -166.46839080459725\n",
      "Episode: 88 \t\t Reward: -170.71931818181778\n",
      "Episode: 89 \t\t Reward: -169.5820224719097\n",
      "Episode: 90 \t\t Reward: -168.09333333333294\n",
      "Episode: 91 \t\t Reward: -166.54175824175786\n",
      "Episode: 92 \t\t Reward: -170.7201086956518\n",
      "Episode: 93 \t\t Reward: -169.16451612903188\n",
      "Episode: 94 \t\t Reward: -168.26276595744642\n",
      "Episode: 95 \t\t Reward: -166.3947368421049\n",
      "Episode: 96 \t\t Reward: -165.2562499999996\n",
      "Episode: 97 \t\t Reward: -168.90721649484496\n",
      "Episode: 98 \t\t Reward: -167.42193877550983\n",
      "Episode: 99 \t\t Reward: -165.96919191919153\n",
      "Episode: 100 \t\t Reward: -170.1514999999996\n",
      "Episode: 101 \t\t Reward: -170.5979999999996\n",
      "Episode: 102 \t\t Reward: -174.9229999999996\n",
      "Episode: 103 \t\t Reward: -175.1644999999996\n",
      "Episode: 104 \t\t Reward: -175.2284999999996\n",
      "Episode: 105 \t\t Reward: -169.36999999999955\n",
      "Episode: 106 \t\t Reward: -169.16349999999954\n",
      "Episode: 107 \t\t Reward: -163.90149999999954\n",
      "Episode: 108 \t\t Reward: -157.80199999999957\n",
      "Episode: 109 \t\t Reward: -163.25899999999953\n",
      "Episode: 110 \t\t Reward: -165.85549999999958\n",
      "Episode: 111 \t\t Reward: -165.6229999999996\n",
      "Episode: 112 \t\t Reward: -166.06349999999958\n",
      "Episode: 113 \t\t Reward: -166.1684999999996\n",
      "Episode: 114 \t\t Reward: -166.4349999999996\n",
      "Episode: 115 \t\t Reward: -166.4839999999996\n",
      "Episode: 116 \t\t Reward: -166.6074999999996\n",
      "Episode: 117 \t\t Reward: -161.5309999999996\n",
      "Episode: 118 \t\t Reward: -166.8579999999996\n",
      "Episode: 119 \t\t Reward: -167.99699999999964\n",
      "Episode: 120 \t\t Reward: -167.25349999999963\n",
      "Episode: 121 \t\t Reward: -172.17099999999965\n",
      "Episode: 122 \t\t Reward: -166.23849999999962\n",
      "Episode: 123 \t\t Reward: -167.07399999999964\n",
      "Episode: 124 \t\t Reward: -165.28849999999963\n",
      "Episode: 125 \t\t Reward: -164.43399999999963\n",
      "Episode: 126 \t\t Reward: -165.06649999999965\n",
      "Episode: 127 \t\t Reward: -164.71549999999968\n",
      "Episode: 128 \t\t Reward: -159.61999999999964\n",
      "Episode: 129 \t\t Reward: -154.70099999999965\n",
      "Episode: 130 \t\t Reward: -154.43199999999965\n",
      "Episode: 131 \t\t Reward: -160.03199999999964\n",
      "Episode: 132 \t\t Reward: -160.06199999999967\n",
      "Episode: 133 \t\t Reward: -164.6764999999996\n",
      "Episode: 134 \t\t Reward: -165.20749999999964\n",
      "Episode: 135 \t\t Reward: -160.03399999999962\n",
      "Episode: 136 \t\t Reward: -161.17649999999966\n",
      "Episode: 137 \t\t Reward: -155.96599999999964\n",
      "Episode: 138 \t\t Reward: -155.81749999999965\n",
      "Episode: 139 \t\t Reward: -155.05549999999963\n",
      "Episode: 140 \t\t Reward: -154.94049999999964\n",
      "Episode: 141 \t\t Reward: -160.55249999999967\n",
      "Episode: 142 \t\t Reward: -155.62749999999963\n",
      "Episode: 143 \t\t Reward: -150.44199999999964\n",
      "Episode: 144 \t\t Reward: -149.9549999999996\n",
      "Episode: 145 \t\t Reward: -156.05699999999965\n",
      "Episode: 146 \t\t Reward: -160.9634999999996\n",
      "Episode: 147 \t\t Reward: -156.20249999999965\n",
      "Episode: 148 \t\t Reward: -150.29049999999958\n",
      "Episode: 149 \t\t Reward: -144.94799999999958\n",
      "Episode: 150 \t\t Reward: -139.71249999999958\n",
      "Episode: 151 \t\t Reward: -141.5274999999996\n",
      "Episode: 152 \t\t Reward: -145.81999999999962\n",
      "Episode: 153 \t\t Reward: -145.37799999999962\n",
      "Episode: 154 \t\t Reward: -139.6359999999996\n",
      "Episode: 155 \t\t Reward: -134.56199999999956\n",
      "Episode: 156 \t\t Reward: -134.19399999999956\n",
      "Episode: 157 \t\t Reward: -128.39999999999955\n",
      "Episode: 158 \t\t Reward: -122.61249999999954\n",
      "Episode: 159 \t\t Reward: -126.92999999999952\n",
      "Episode: 160 \t\t Reward: -132.00849999999954\n",
      "Episode: 161 \t\t Reward: -132.40649999999954\n",
      "Episode: 162 \t\t Reward: -131.94549999999953\n",
      "Episode: 163 \t\t Reward: -132.87199999999956\n",
      "Episode: 164 \t\t Reward: -132.28649999999953\n",
      "Episode: 165 \t\t Reward: -137.33199999999954\n",
      "Episode: 166 \t\t Reward: -142.85299999999958\n",
      "Episode: 167 \t\t Reward: -148.7084999999996\n",
      "Episode: 168 \t\t Reward: -148.7524999999996\n",
      "Episode: 169 \t\t Reward: -148.21899999999962\n",
      "Episode: 170 \t\t Reward: -148.5099999999996\n",
      "Episode: 171 \t\t Reward: -148.3709999999996\n",
      "Episode: 172 \t\t Reward: -153.7334999999996\n",
      "Episode: 173 \t\t Reward: -152.80549999999957\n",
      "Episode: 174 \t\t Reward: -147.29099999999954\n",
      "Episode: 175 \t\t Reward: -145.52249999999955\n",
      "Episode: 176 \t\t Reward: -144.2144999999995\n",
      "Episode: 177 \t\t Reward: -139.2419999999995\n",
      "Episode: 178 \t\t Reward: -139.33449999999948\n",
      "Episode: 179 \t\t Reward: -139.15899999999948\n",
      "Episode: 180 \t\t Reward: -139.38499999999948\n",
      "Episode: 181 \t\t Reward: -134.6244999999995\n",
      "Episode: 182 \t\t Reward: -134.29099999999949\n",
      "Episode: 183 \t\t Reward: -133.05699999999948\n",
      "Episode: 184 \t\t Reward: -134.5004999999995\n",
      "Episode: 185 \t\t Reward: -130.2374999999995\n",
      "Episode: 186 \t\t Reward: -131.19649999999953\n",
      "Episode: 187 \t\t Reward: -130.79099999999954\n",
      "Episode: 188 \t\t Reward: -125.1444999999995\n",
      "Episode: 189 \t\t Reward: -130.1244999999995\n",
      "Episode: 190 \t\t Reward: -135.4734999999995\n",
      "Episode: 191 \t\t Reward: -135.7899999999995\n",
      "Episode: 192 \t\t Reward: -130.9424999999995\n",
      "Episode: 193 \t\t Reward: -131.2384999999995\n",
      "Episode: 194 \t\t Reward: -130.25249999999951\n",
      "Episode: 195 \t\t Reward: -130.4514999999995\n",
      "Episode: 196 \t\t Reward: -129.87899999999948\n",
      "Episode: 197 \t\t Reward: -124.62499999999947\n",
      "Episode: 198 \t\t Reward: -124.85149999999949\n",
      "Episode: 199 \t\t Reward: -124.61399999999949\n",
      "Episode: 200 \t\t Reward: -118.31449999999948\n",
      "Episode: 201 \t\t Reward: -118.14699999999947\n",
      "Episode: 202 \t\t Reward: -113.6439999999995\n",
      "Episode: 203 \t\t Reward: -114.37749999999949\n",
      "Episode: 204 \t\t Reward: -113.12999999999946\n",
      "Episode: 205 \t\t Reward: -118.12549999999946\n",
      "Episode: 206 \t\t Reward: -119.12199999999947\n",
      "Episode: 207 \t\t Reward: -118.95249999999946\n",
      "Episode: 208 \t\t Reward: -125.10449999999946\n",
      "Episode: 209 \t\t Reward: -117.58399999999946\n",
      "Episode: 210 \t\t Reward: -116.74999999999946\n",
      "Episode: 211 \t\t Reward: -117.06349999999944\n",
      "Episode: 212 \t\t Reward: -116.57699999999946\n",
      "Episode: 213 \t\t Reward: -116.06749999999946\n",
      "Episode: 214 \t\t Reward: -121.52799999999945\n",
      "Episode: 215 \t\t Reward: -121.20049999999942\n",
      "Episode: 216 \t\t Reward: -121.59349999999944\n",
      "Episode: 217 \t\t Reward: -120.50999999999944\n",
      "Episode: 218 \t\t Reward: -114.38149999999938\n",
      "Episode: 219 \t\t Reward: -114.0254999999994\n",
      "Episode: 220 \t\t Reward: -113.6229999999994\n",
      "Episode: 221 \t\t Reward: -107.04199999999935\n",
      "Episode: 222 \t\t Reward: -112.58799999999937\n",
      "Episode: 223 \t\t Reward: -117.22399999999936\n",
      "Episode: 224 \t\t Reward: -119.16149999999938\n",
      "Episode: 225 \t\t Reward: -119.36599999999939\n",
      "Episode: 226 \t\t Reward: -118.9239999999994\n",
      "Episode: 227 \t\t Reward: -118.6724999999994\n",
      "Episode: 228 \t\t Reward: -119.1514999999994\n",
      "Episode: 229 \t\t Reward: -119.0114999999994\n",
      "Episode: 230 \t\t Reward: -118.81149999999938\n",
      "Episode: 231 \t\t Reward: -113.65849999999936\n",
      "Episode: 232 \t\t Reward: -112.29349999999936\n",
      "Episode: 233 \t\t Reward: -107.48549999999933\n",
      "Episode: 234 \t\t Reward: -107.02649999999932\n",
      "Episode: 235 \t\t Reward: -112.43699999999936\n",
      "Episode: 236 \t\t Reward: -111.97249999999933\n",
      "Episode: 237 \t\t Reward: -112.56749999999933\n",
      "Episode: 238 \t\t Reward: -111.81999999999933\n",
      "Episode: 239 \t\t Reward: -112.25299999999933\n",
      "Episode: 240 \t\t Reward: -112.99849999999935\n",
      "Episode: 241 \t\t Reward: -108.46399999999934\n",
      "Episode: 242 \t\t Reward: -107.50999999999932\n",
      "Episode: 243 \t\t Reward: -112.47449999999934\n",
      "Episode: 244 \t\t Reward: -112.72199999999934\n",
      "Episode: 245 \t\t Reward: -107.0794999999993\n",
      "Episode: 246 \t\t Reward: -102.17299999999932\n",
      "Episode: 247 \t\t Reward: -102.20499999999933\n",
      "Episode: 248 \t\t Reward: -100.87549999999932\n",
      "Episode: 249 \t\t Reward: -101.97499999999937\n",
      "Episode: 250 \t\t Reward: -106.78849999999935\n",
      "Episode: 251 \t\t Reward: -106.75199999999936\n",
      "Episode: 252 \t\t Reward: -99.75449999999934\n",
      "Episode: 253 \t\t Reward: -99.70049999999932\n",
      "Episode: 254 \t\t Reward: -100.44949999999935\n",
      "Episode: 255 \t\t Reward: -100.65849999999936\n",
      "Episode: 256 \t\t Reward: -99.78149999999934\n",
      "Episode: 257 \t\t Reward: -100.41399999999936\n",
      "Episode: 258 \t\t Reward: -100.29499999999935\n",
      "Episode: 259 \t\t Reward: -93.46899999999931\n",
      "Episode: 260 \t\t Reward: -87.8604999999993\n",
      "Episode: 261 \t\t Reward: -87.5734999999993\n",
      "Episode: 262 \t\t Reward: -87.7164999999993\n",
      "Episode: 263 \t\t Reward: -87.73099999999931\n",
      "Episode: 264 \t\t Reward: -83.16149999999931\n",
      "Episode: 265 \t\t Reward: -82.25099999999928\n",
      "Episode: 266 \t\t Reward: -76.51999999999926\n",
      "Episode: 267 \t\t Reward: -70.73699999999926\n",
      "Episode: 268 \t\t Reward: -71.10699999999925\n",
      "Episode: 269 \t\t Reward: -71.09599999999925\n",
      "Episode: 270 \t\t Reward: -69.98149999999923\n",
      "Episode: 271 \t\t Reward: -70.09599999999924\n",
      "Episode: 272 \t\t Reward: -63.97199999999921\n",
      "Episode: 273 \t\t Reward: -64.94149999999925\n",
      "Episode: 274 \t\t Reward: -65.39799999999927\n",
      "Episode: 275 \t\t Reward: -67.90449999999929\n",
      "Episode: 276 \t\t Reward: -68.96699999999933\n",
      "Episode: 277 \t\t Reward: -73.59649999999932\n",
      "Episode: 278 \t\t Reward: -74.32099999999934\n",
      "Episode: 279 \t\t Reward: -68.28449999999933\n",
      "Episode: 280 \t\t Reward: -68.12599999999934\n",
      "Episode: 281 \t\t Reward: -71.7529999999993\n",
      "Episode: 282 \t\t Reward: -72.2779999999993\n",
      "Episode: 283 \t\t Reward: -77.74549999999931\n",
      "Episode: 284 \t\t Reward: -77.0499999999993\n",
      "Episode: 285 \t\t Reward: -76.5324999999993\n",
      "Episode: 286 \t\t Reward: -74.37149999999924\n",
      "Episode: 287 \t\t Reward: -74.38299999999927\n",
      "Episode: 288 \t\t Reward: -75.46949999999929\n",
      "Episode: 289 \t\t Reward: -75.30599999999929\n",
      "Episode: 290 \t\t Reward: -69.8859999999993\n",
      "Episode: 291 \t\t Reward: -75.23499999999927\n",
      "Episode: 292 \t\t Reward: -74.31999999999927\n",
      "Episode: 293 \t\t Reward: -74.16949999999925\n",
      "Episode: 294 \t\t Reward: -79.63599999999927\n",
      "Episode: 295 \t\t Reward: -85.02899999999929\n",
      "Episode: 296 \t\t Reward: -90.3549999999993\n",
      "Episode: 297 \t\t Reward: -95.3734999999993\n",
      "Episode: 298 \t\t Reward: -94.47299999999926\n",
      "Episode: 299 \t\t Reward: -100.34199999999926\n",
      "Episode: 300 \t\t Reward: -100.54049999999927\n",
      "Episode: 301 \t\t Reward: -100.28049999999926\n",
      "Episode: 302 \t\t Reward: -100.69799999999928\n",
      "Episode: 303 \t\t Reward: -100.24199999999928\n",
      "Episode: 304 \t\t Reward: -101.6699999999993\n",
      "Episode: 305 \t\t Reward: -97.42999999999931\n",
      "Episode: 306 \t\t Reward: -97.39149999999931\n",
      "Episode: 307 \t\t Reward: -97.69699999999929\n",
      "Episode: 308 \t\t Reward: -91.7894999999993\n",
      "Episode: 309 \t\t Reward: -91.7794999999993\n",
      "Episode: 310 \t\t Reward: -92.2419999999993\n",
      "Episode: 311 \t\t Reward: -92.0034999999993\n",
      "Episode: 312 \t\t Reward: -92.26549999999932\n",
      "Episode: 313 \t\t Reward: -93.15199999999935\n",
      "Episode: 314 \t\t Reward: -92.97249999999934\n",
      "Episode: 315 \t\t Reward: -93.36499999999934\n",
      "Episode: 316 \t\t Reward: -92.62999999999933\n",
      "Episode: 317 \t\t Reward: -92.97249999999931\n",
      "Episode: 318 \t\t Reward: -92.95749999999931\n",
      "Episode: 319 \t\t Reward: -93.21799999999932\n",
      "Episode: 320 \t\t Reward: -93.49199999999932\n",
      "Episode: 321 \t\t Reward: -95.04499999999935\n",
      "Episode: 322 \t\t Reward: -89.91149999999932\n",
      "Episode: 323 \t\t Reward: -90.78199999999933\n",
      "Episode: 324 \t\t Reward: -91.35449999999935\n",
      "Episode: 325 \t\t Reward: -91.47199999999935\n",
      "Episode: 326 \t\t Reward: -91.72899999999937\n",
      "Episode: 327 \t\t Reward: -92.29849999999938\n",
      "Episode: 328 \t\t Reward: -96.29949999999938\n",
      "Episode: 329 \t\t Reward: -101.47149999999937\n",
      "Episode: 330 \t\t Reward: -101.12449999999936\n",
      "Episode: 331 \t\t Reward: -105.84249999999936\n",
      "Episode: 332 \t\t Reward: -112.3624999999994\n",
      "Episode: 333 \t\t Reward: -111.99749999999938\n",
      "Episode: 334 \t\t Reward: -112.23349999999938\n",
      "Episode: 335 \t\t Reward: -106.32249999999938\n",
      "Episode: 336 \t\t Reward: -105.80049999999936\n",
      "Episode: 337 \t\t Reward: -110.62749999999939\n",
      "Episode: 338 \t\t Reward: -110.34999999999937\n",
      "Episode: 339 \t\t Reward: -110.27199999999935\n",
      "Episode: 340 \t\t Reward: -114.65199999999935\n",
      "Episode: 341 \t\t Reward: -113.40399999999933\n",
      "Episode: 342 \t\t Reward: -119.39499999999933\n",
      "Episode: 343 \t\t Reward: -119.35799999999932\n",
      "Episode: 344 \t\t Reward: -120.00099999999935\n",
      "Episode: 345 \t\t Reward: -120.57999999999934\n",
      "Episode: 346 \t\t Reward: -120.94099999999935\n",
      "Episode: 347 \t\t Reward: -121.84349999999937\n",
      "Episode: 348 \t\t Reward: -123.60749999999942\n",
      "Episode: 349 \t\t Reward: -122.9774999999994\n",
      "Episode: 350 \t\t Reward: -118.58999999999938\n",
      "Episode: 351 \t\t Reward: -118.7949999999994\n",
      "Episode: 352 \t\t Reward: -125.2624999999994\n",
      "Episode: 353 \t\t Reward: -131.1569999999994\n",
      "Episode: 354 \t\t Reward: -130.9579999999994\n",
      "Episode: 355 \t\t Reward: -130.7154999999994\n",
      "Episode: 356 \t\t Reward: -132.0814999999994\n",
      "Episode: 357 \t\t Reward: -136.63599999999943\n",
      "Episode: 358 \t\t Reward: -137.1814999999994\n",
      "Episode: 359 \t\t Reward: -139.15099999999947\n",
      "Episode: 360 \t\t Reward: -145.03799999999947\n",
      "Episode: 361 \t\t Reward: -149.80199999999948\n",
      "Episode: 362 \t\t Reward: -149.05499999999947\n",
      "Episode: 363 \t\t Reward: -148.99349999999944\n",
      "Episode: 364 \t\t Reward: -148.68699999999944\n",
      "Episode: 365 \t\t Reward: -143.51499999999945\n",
      "Episode: 366 \t\t Reward: -144.11999999999946\n",
      "Episode: 367 \t\t Reward: -143.63349999999946\n",
      "Episode: 368 \t\t Reward: -143.59149999999943\n",
      "Episode: 369 \t\t Reward: -143.75649999999945\n",
      "Episode: 370 \t\t Reward: -143.96249999999947\n",
      "Episode: 371 \t\t Reward: -144.12799999999945\n",
      "Episode: 372 \t\t Reward: -144.76099999999946\n",
      "Episode: 373 \t\t Reward: -144.8904999999995\n",
      "Episode: 374 \t\t Reward: -144.2569999999995\n",
      "Episode: 375 \t\t Reward: -143.59249999999946\n",
      "Episode: 376 \t\t Reward: -143.04549999999944\n",
      "Episode: 377 \t\t Reward: -143.13049999999944\n",
      "Episode: 378 \t\t Reward: -143.16899999999944\n",
      "Episode: 379 \t\t Reward: -143.59349999999947\n",
      "Episode: 380 \t\t Reward: -143.47899999999944\n",
      "Episode: 381 \t\t Reward: -139.21599999999947\n",
      "Episode: 382 \t\t Reward: -138.61649999999946\n",
      "Episode: 383 \t\t Reward: -133.12599999999946\n",
      "Episode: 384 \t\t Reward: -137.87399999999946\n",
      "Episode: 385 \t\t Reward: -143.45499999999944\n",
      "Episode: 386 \t\t Reward: -151.18299999999948\n",
      "Episode: 387 \t\t Reward: -151.54349999999948\n",
      "Episode: 388 \t\t Reward: -155.7049999999995\n",
      "Episode: 389 \t\t Reward: -150.58699999999948\n",
      "Episode: 390 \t\t Reward: -150.4019999999995\n",
      "Episode: 391 \t\t Reward: -150.0249999999995\n",
      "Episode: 392 \t\t Reward: -150.6749999999995\n",
      "Episode: 393 \t\t Reward: -150.7429999999995\n",
      "Episode: 394 \t\t Reward: -150.5384999999995\n",
      "Episode: 395 \t\t Reward: -145.16649999999947\n",
      "Episode: 396 \t\t Reward: -140.09599999999946\n",
      "Episode: 397 \t\t Reward: -135.81999999999948\n",
      "Episode: 398 \t\t Reward: -135.71599999999947\n",
      "Episode: 399 \t\t Reward: -129.87399999999946\n",
      "Episode: 400 \t\t Reward: -130.2264999999995\n",
      "Episode: 401 \t\t Reward: -135.9704999999995\n",
      "Episode: 402 \t\t Reward: -135.32599999999945\n",
      "Episode: 403 \t\t Reward: -139.55199999999945\n",
      "Episode: 404 \t\t Reward: -138.22499999999943\n",
      "Episode: 405 \t\t Reward: -137.0174999999994\n",
      "Episode: 406 \t\t Reward: -136.62449999999941\n",
      "Episode: 407 \t\t Reward: -137.53149999999943\n",
      "Episode: 408 \t\t Reward: -137.60199999999944\n",
      "Episode: 409 \t\t Reward: -138.91949999999946\n",
      "Episode: 410 \t\t Reward: -136.69449999999944\n",
      "Episode: 411 \t\t Reward: -137.03049999999942\n",
      "Episode: 412 \t\t Reward: -137.1719999999994\n",
      "Episode: 413 \t\t Reward: -137.25299999999942\n",
      "Episode: 414 \t\t Reward: -131.64999999999935\n",
      "Episode: 415 \t\t Reward: -136.48549999999938\n",
      "Episode: 416 \t\t Reward: -137.11949999999942\n",
      "Episode: 417 \t\t Reward: -137.93549999999942\n",
      "Episode: 418 \t\t Reward: -144.13399999999945\n",
      "Episode: 419 \t\t Reward: -144.17399999999944\n",
      "Episode: 420 \t\t Reward: -144.63899999999944\n",
      "Episode: 421 \t\t Reward: -145.61099999999948\n",
      "Episode: 422 \t\t Reward: -151.01799999999946\n",
      "Episode: 423 \t\t Reward: -145.3809999999995\n",
      "Episode: 424 \t\t Reward: -142.85099999999943\n",
      "Episode: 425 \t\t Reward: -143.2784999999994\n",
      "Episode: 426 \t\t Reward: -148.78849999999943\n",
      "Episode: 427 \t\t Reward: -148.87049999999945\n",
      "Episode: 428 \t\t Reward: -149.59749999999946\n",
      "Episode: 429 \t\t Reward: -144.57999999999944\n",
      "Episode: 430 \t\t Reward: -144.42149999999944\n",
      "Episode: 431 \t\t Reward: -144.52449999999945\n",
      "Episode: 432 \t\t Reward: -138.80899999999946\n",
      "Episode: 433 \t\t Reward: -139.53049999999948\n",
      "Episode: 434 \t\t Reward: -139.46849999999947\n",
      "Episode: 435 \t\t Reward: -140.2824999999995\n",
      "Episode: 436 \t\t Reward: -141.35949999999954\n",
      "Episode: 437 \t\t Reward: -136.5014999999995\n",
      "Episode: 438 \t\t Reward: -141.83299999999952\n",
      "Episode: 439 \t\t Reward: -147.30649999999955\n",
      "Episode: 440 \t\t Reward: -141.5104999999995\n",
      "Episode: 441 \t\t Reward: -142.12649999999954\n",
      "Episode: 442 \t\t Reward: -136.42249999999953\n",
      "Episode: 443 \t\t Reward: -130.93349999999953\n",
      "Episode: 444 \t\t Reward: -130.41549999999953\n",
      "Episode: 445 \t\t Reward: -135.54749999999953\n",
      "Episode: 446 \t\t Reward: -135.8429999999995\n",
      "Episode: 447 \t\t Reward: -141.19649999999953\n",
      "Episode: 448 \t\t Reward: -146.86499999999953\n",
      "Episode: 449 \t\t Reward: -146.78799999999953\n",
      "Episode: 450 \t\t Reward: -151.83749999999952\n",
      "Episode: 451 \t\t Reward: -151.6324999999995\n",
      "Episode: 452 \t\t Reward: -147.02249999999952\n",
      "Episode: 453 \t\t Reward: -141.65899999999954\n",
      "Episode: 454 \t\t Reward: -142.3764999999995\n",
      "Episode: 455 \t\t Reward: -142.52849999999953\n",
      "Episode: 456 \t\t Reward: -142.54149999999953\n",
      "Episode: 457 \t\t Reward: -137.73349999999954\n",
      "Episode: 458 \t\t Reward: -137.33699999999956\n",
      "Episode: 459 \t\t Reward: -143.00949999999955\n",
      "Episode: 460 \t\t Reward: -136.91899999999953\n",
      "Episode: 461 \t\t Reward: -131.99699999999956\n",
      "Episode: 462 \t\t Reward: -132.12099999999955\n",
      "Episode: 463 \t\t Reward: -130.8094999999995\n",
      "Episode: 464 \t\t Reward: -130.9679999999995\n",
      "Episode: 465 \t\t Reward: -131.75799999999953\n",
      "Episode: 466 \t\t Reward: -137.10199999999952\n",
      "Episode: 467 \t\t Reward: -143.01849999999953\n",
      "Episode: 468 \t\t Reward: -142.13549999999952\n",
      "Episode: 469 \t\t Reward: -142.2624999999995\n",
      "Episode: 470 \t\t Reward: -142.50099999999955\n",
      "Episode: 471 \t\t Reward: -142.68899999999954\n",
      "Episode: 472 \t\t Reward: -142.81249999999952\n",
      "Episode: 473 \t\t Reward: -142.81549999999953\n",
      "Episode: 474 \t\t Reward: -143.13999999999953\n"
     ]
    },
    {
     "ename": "OutOfMemoryError",
     "evalue": "CUDA out of memory. Tried to allocate 2.00 MiB (GPU 0; 8.00 GiB total capacity; 7.30 GiB already allocated; 0 bytes free; 7.31 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mOutOfMemoryError\u001b[0m                          Traceback (most recent call last)",
      "\u001b[1;32mc:\\Users\\lynch\\School\\intership\\iecon_2023_rl_reward_shaping\\LoadModelNTest.ipynb Cell 21\u001b[0m in \u001b[0;36m<cell line: 18>\u001b[1;34m()\u001b[0m\n\u001b[0;32m     <a href='vscode-notebook-cell:/c%3A/Users/lynch/School/intership/iecon_2023_rl_reward_shaping/LoadModelNTest.ipynb#X25sZmlsZQ%3D%3D?line=24'>25</a>\u001b[0m Ep_bump \u001b[39m=\u001b[39m[]\n\u001b[0;32m     <a href='vscode-notebook-cell:/c%3A/Users/lynch/School/intership/iecon_2023_rl_reward_shaping/LoadModelNTest.ipynb#X25sZmlsZQ%3D%3D?line=25'>26</a>\u001b[0m \u001b[39mfor\u001b[39;00m t \u001b[39min\u001b[39;00m \u001b[39mrange\u001b[39m(\u001b[39m1\u001b[39m, \u001b[39m1500\u001b[39m):\n\u001b[1;32m---> <a href='vscode-notebook-cell:/c%3A/Users/lynch/School/intership/iecon_2023_rl_reward_shaping/LoadModelNTest.ipynb#X25sZmlsZQ%3D%3D?line=26'>27</a>\u001b[0m     action \u001b[39m=\u001b[39m agent_PPO\u001b[39m.\u001b[39;49mselect_action(state)\n\u001b[0;32m     <a href='vscode-notebook-cell:/c%3A/Users/lynch/School/intership/iecon_2023_rl_reward_shaping/LoadModelNTest.ipynb#X25sZmlsZQ%3D%3D?line=27'>28</a>\u001b[0m     ac\u001b[39m.\u001b[39mappend(action)\n\u001b[0;32m     <a href='vscode-notebook-cell:/c%3A/Users/lynch/School/intership/iecon_2023_rl_reward_shaping/LoadModelNTest.ipynb#X25sZmlsZQ%3D%3D?line=28'>29</a>\u001b[0m     state, reward, done, bumps \u001b[39m=\u001b[39m env\u001b[39m.\u001b[39mstep(action)\n",
      "File \u001b[1;32mc:\\Users\\lynch\\School\\intership\\iecon_2023_rl_reward_shaping\\agent\\ppoagent.py:188\u001b[0m, in \u001b[0;36mPPO.select_action\u001b[1;34m(self, state)\u001b[0m\n\u001b[0;32m    186\u001b[0m \u001b[39melse\u001b[39;00m:\n\u001b[0;32m    187\u001b[0m     \u001b[39mwith\u001b[39;00m torch\u001b[39m.\u001b[39mno_grad():\n\u001b[1;32m--> 188\u001b[0m         state \u001b[39m=\u001b[39m torch\u001b[39m.\u001b[39;49mFloatTensor(state)\u001b[39m.\u001b[39;49mto(device)\n\u001b[0;32m    189\u001b[0m         action, action_logprob, state_val \u001b[39m=\u001b[39m \u001b[39mself\u001b[39m\u001b[39m.\u001b[39mpolicy_old\u001b[39m.\u001b[39mact(state)\n\u001b[0;32m    191\u001b[0m     \u001b[39mself\u001b[39m\u001b[39m.\u001b[39mbuffer\u001b[39m.\u001b[39mstates\u001b[39m.\u001b[39mappend(state)\n",
      "\u001b[1;31mOutOfMemoryError\u001b[0m: CUDA out of memory. Tried to allocate 2.00 MiB (GPU 0; 8.00 GiB total capacity; 7.30 GiB already allocated; 0 bytes free; 7.31 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation.  See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF"
     ]
    }
   ],
   "source": [
    "from collections import deque\n",
    "import numpy as np\n",
    "from utility import get_unique_numbers\n",
    "import time\n",
    "\n",
    "test_running_reward = 0\n",
    "# total_test_episodes = 3\n",
    "ac =[]\n",
    "start = time.time()\n",
    "runn_ep =[]\n",
    "scores =[]\n",
    "scores_deque = deque(maxlen=100)\n",
    "avg_scores =[]\n",
    "Bmp =[]\n",
    "T_Bmp = []\n",
    "if p_model.is_dir():\n",
    "    nb_name = 'PPO_Test_log.txt'\n",
    "with open(p_model.joinpath(nb_name), \"w\") as f:\n",
    "\n",
    "    for ep in range(1, total_test_episodes+1):\n",
    "        start_ep = time.time()\n",
    "        ep_reward = 0\n",
    "        runn_ep.append(ep)\n",
    "        state = env.reset()\n",
    "        Ep_bump =[]\n",
    "        for t in range(1, 1500):\n",
    "            action = agent_PPO.select_action(state)\n",
    "            ac.append(action)\n",
    "            state, reward, done, bumps = env.step(action)\n",
    "            Ep_bump.append(bumps)\n",
    "            T_Bmp.append(bumps)\n",
    "            ep_reward += reward\n",
    "            # env.render('rgb_array') # uncommand to display -- warning: expensive resources.\n",
    "            if done:\n",
    "                break\n",
    "        Bmp.append(Ep_bump)\n",
    "        scores_deque.append(ep_reward)\n",
    "        avg_scores.append(np.mean(scores_deque))\n",
    "        end_ep = time.time()\n",
    "        hours_ep, rem_ep = divmod(end_ep-start_ep, 3600)\n",
    "        minutes_ep, seconds_ep = divmod(rem_ep, 60)\n",
    "\n",
    "        f.write(\"Episode{} (survive time): {:0>2}:{:0>2}:{:05.2f}\\n\".format(ep,int(hours_ep),int(minutes_ep),seconds_ep))\n",
    "        test_running_reward +=  ep_reward\n",
    "\n",
    "        # print('Episode: {} \\t\\t Reward: {}'.format(ep, round(ep_reward, 2)))\n",
    "        print('Episode: {} \\t\\t Reward: {}'.format(ep, avg_scores[ep-1]))\n",
    "\n",
    "        scores.append(round(ep_reward, 2))\n",
    "        ep_reward = 0\n",
    "    env.close()\n",
    "\n",
    "    end = time.time()\n",
    "    hours, rem = divmod(end-start, 3600)\n",
    "    minutes, seconds = divmod(rem, 60)\n",
    "    print(\"Total test time: {:0>2}:{:0>2}:{:05.2f}\".format(int(hours),int(minutes),seconds))\n",
    "    \n",
    "    # count the bump instances.\n",
    "    count_list_PPO = []\n",
    "    for i in range(len(Bmp)):\n",
    "        count_list_PPO.append(count_impulses(Bmp[i]))\n",
    "    data_PPO = {'Bump_instance': count_list_PPO}\n",
    "    Bump_track = pd.DataFrame(data_PPO)\n",
    "    Tcount_PPO = count_edgezone(T_Bmp,gap_threshold)\n",
    "\n",
    "\n",
    "    f.write(\"Total Test time (survive time): {:0>2}:{:0>2}:{:05.2f}\\n\".format(int(hours),int(minutes),seconds))\n",
    "    f.write('Toal epoch:%s\\n'%total_test_episodes)\n",
    "    f.write(\"Bump instances of total: %s\\n\"%Tcount_PPO)\n",
    "    f.write(\"Bump to edge zone %s\\n\"%count_edgezone(T_Bmp, gap_threshold))\n",
    "    f.close()\n",
    "\n",
    "PPO_avg = avg_scores\n",
    "PPO_scores = scores\n",
    "PPO_ac = ac\n",
    "\n",
    "csv_name = f\"PPO_default\"\n",
    "# score_save(scores,avg_scores,file_name=csv_name,path=p_model,print_=False) # save score table as csv file\n",
    "score_action_save(PPO_ac,PPO_scores,PPO_avg,file_name=csv_name,path=p_model,print_=False) # save score table as csv file\n",
    "count_impulse_file = f\"PPO_default_count\"\n",
    "Bump_track.to_csv((p_model/f\"{count_impulse_file}.csv\")) # save bump instances as csv file\n",
    "\n",
    "# count the bump instances.\n",
    "count_list_PPO = []\n",
    "for i in range(len(Bmp)):\n",
    "    count_list_PPO.append(count_impulses(Bmp[i]))\n",
    "data_PPO = {'Bump_instance': count_list_PPO}\n",
    "Bump_track_PPO = pd.DataFrame(data_PPO)\n",
    "\n",
    "# fig = plt.figure()\n",
    "# plt.plot(T_Bmp)\n",
    "# plt.title(f'Total bump instances:{count_impulses(T_Bmp)}')\n",
    "\n",
    "# # plot the bump instances.\n",
    "# for i in range(5):\n",
    "#     plt.subplot(2, 5, i+1)\n",
    "#     plt.plot(Bmp[i])\n",
    "#     plt.title(f'Ct_:{count_list[i]}')\n",
    "#     plt.subplot(2, 5, i+6)\n",
    "#     plt.plot(Bmp[i+5])\n",
    "#     plt.title(f'Ct_:{count_list[i+5]}')\n",
    "\n",
    "# fig = plt.figure()\n",
    "# plt.plot(T_Bmp)\n",
    "# plt.title(f'Total bump instances:{Tcount_PPO}')\n",
    "\n",
    "\n",
    "fig = plt.figure()\n",
    "ax = fig.add_subplot(111)\n",
    "plt.plot(avg_scores)\n",
    "# plt.plot(df['Averagescores'])\n",
    "plt.ylabel('Scores')\n",
    "plt.xlabel(f'Episode {len(avg_scores)}')\n",
    "plt.show()\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "\n",
    "action_list = [0,1,2,3,4]\n",
    "h=[]\n",
    "for i in action_list:\n",
    "    h.append(ac.count(i))\n",
    "bar_contain = plt.bar(action_list,h)\n",
    "plt.xlabel('Action items')\n",
    "plt.ylabel('Number of action item used during trial')\n",
    "plt.bar_label(bar_contain)\n",
    "plt.show()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## DQN agent test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Episode: 1 \t\t Reward: 0.15\n",
      "Episode: 2 \t\t Reward: -13.850000000000062\n",
      "Episode: 3 \t\t Reward: -18.516666666666747\n",
      "Episode: 4 \t\t Reward: -20.850000000000094\n",
      "Episode: 5 \t\t Reward: -22.2800000000001\n",
      "Total test time: 00:00:07.10\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from collections import deque\n",
    "import numpy as np\n",
    "from utility import get_unique_numbers\n",
    "import time\n",
    "\n",
    "test_running_reward = 0\n",
    "\n",
    "ac =[]\n",
    "start = time.time()\n",
    "runn_ep =[]\n",
    "scores =[]\n",
    "scores_deque = deque(maxlen=100)\n",
    "avg_scores =[]\n",
    "Bmp = []\n",
    "T_Bmp = []\n",
    "bumps = 0\n",
    "\n",
    "if p_model.is_dir():\n",
    "    nb_name = 'DQN_Test_log.txt'\n",
    "with open(p_model.joinpath(nb_name), \"w\") as f:\n",
    "\n",
    "    for ep in range(1, total_test_episodes+1):\n",
    "        start_ep = time.time()\n",
    "        ep_reward = 0\n",
    "        runn_ep.append(ep)\n",
    "        state = env.reset()\n",
    "        Ep_bmp = []\n",
    "\n",
    "        for t in range(1, 1500):\n",
    "            action = agent_DQN.act(state)\n",
    "            ac.append(action)\n",
    "            state, reward, done, bumps = env.step(action)\n",
    "\n",
    "            Ep_bmp.append(bumps)\n",
    "            T_Bmp.append(bumps)\n",
    "            ep_reward += reward\n",
    "            # env.render('rgb_array') # uncommand to display -- warning: expensive resources.\n",
    "            if done:\n",
    "                break\n",
    "        \n",
    "        Bmp.append(Ep_bmp)\n",
    "        scores_deque.append(ep_reward)\n",
    "        avg_scores.append(np.mean(scores_deque))\n",
    "        end_ep = time.time()\n",
    "        hours_ep, rem_ep = divmod(end_ep-start_ep, 3600)\n",
    "        minutes_ep, seconds_ep = divmod(rem_ep, 60)\n",
    "\n",
    "        f.write(\"Episode{} (survive time): {:0>2}:{:0>2}:{:05.2f}\\n\".format(ep,int(hours_ep),int(minutes_ep),seconds_ep))\n",
    "        test_running_reward +=  ep_reward\n",
    "\n",
    "        print('Episode: {} \\t\\t Reward: {}'.format(ep, avg_scores[ep-1]))\n",
    "\n",
    "        scores.append(round(ep_reward, 2))\n",
    "        ep_reward = 0\n",
    "    env.close()\n",
    "\n",
    "    end = time.time()\n",
    "    hours, rem = divmod(end-start, 3600)\n",
    "    minutes, seconds = divmod(rem, 60)\n",
    "    print(\"Total test time: {:0>2}:{:0>2}:{:05.2f}\".format(int(hours),int(minutes),seconds))\n",
    "    \n",
    "    # count the bump instances.\n",
    "    count_list_DQN = []\n",
    "    for i in range(len(Bmp)):\n",
    "        count_list_DQN.append(count_impulses(Bmp[i]))\n",
    "    data_dqn = {'Bump_instance': count_list_DQN}\n",
    "    Bump_track = pd.DataFrame(data_dqn)\n",
    "    Tcount_DQN = count_edgezone(T_Bmp,gap_threshold)\n",
    "\n",
    "    f.write(\"Total Test time (survive time): {:0>2}:{:0>2}:{:05.2f}\\n\".format(int(hours),int(minutes),seconds))\n",
    "    f.write('Toal epoch:%s\\n'%total_test_episodes)\n",
    "    f.write(\"Bump instances of total: %s\\n\"%Tcount_DQN)\n",
    "    f.write(\"Bump to edge zone %s\\n\"%count_edgezone(T_Bmp, gap_threshold))\n",
    "    f.close()\n",
    "\n",
    "DQN_avg = avg_scores\n",
    "DQN_scores = scores\n",
    "DQN_ac = ac\n",
    "\n",
    "csv_name = f\"DQN_reshape\"\n",
    "# score_save(scores,avg_scores,file_name=csv_name,path=p_model,print_=False) # save score table as csv file\n",
    "score_action_save(DQN_ac,DQN_scores,DQN_avg,file_name=csv_name,path=p_model,print_=False) # save score table as csv file\n",
    "count_impulse_file = f\"DQN_default_count\"\n",
    "Bump_track.to_csv((p_model/f\"{count_impulse_file}.csv\")) # save bump instances as csv file\n",
    "\n",
    "\n",
    "fig = plt.figure()\n",
    "ax = fig.add_subplot(111)\n",
    "plt.plot(avg_scores)\n",
    "# plt.plot(df['Averagescores'])\n",
    "plt.ylabel('Scores')\n",
    "plt.xlabel(f'Episode {len(avg_scores)}')\n",
    "plt.show()\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "\n",
    "action_list = [0,1,2,3,4]\n",
    "h=[]\n",
    "for i in action_list:\n",
    "    h.append(ac.count(i))\n",
    "bar_contain = plt.bar(action_list,h)\n",
    "plt.xlabel('Action items')\n",
    "plt.ylabel('Number of action item used during trial')\n",
    "plt.bar_label(bar_contain)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "DDPG_ac = pd.Series(DDPG_ac, name='DDPG_ac')\n",
    "DDPG_scores = pd.Series(DDPG_scores, name='DDPG_scores')\n",
    "DDPG_average_score=pd.Series(DDPG_avg,name='DDPG_avg')\n",
    "\n",
    "PPO_ac = pd.Series(PPO_ac, name='PPO_ac')\n",
    "PPO_scores = pd.Series(PPO_scores, name='PPO_scores')\n",
    "PPO_average_score=pd.Series(PPO_avg,name='PPO_avg')\n",
    "\n",
    "DQN_ac = pd.Series(DQN_ac, name='DQN_ac')\n",
    "DQN_scores = pd.Series(DQN_scores, name='DQN_scores')\n",
    "DQN_average_score=pd.Series(DQN_avg,name='DQN_avg')\n",
    "\n",
    "\n",
    "df = pd.concat([DQN_ac,DQN_scores, DQN_average_score, DDPG_ac,DDPG_scores,DDPG_average_score,PPO_ac,PPO_scores,PPO_average_score],axis=1)\n",
    "\n",
    "fig = plt.figure()\n",
    "ax = fig.add_subplot(111)\n",
    "plt.plot(df['DQN_avg'], 'r-.',label='DQN Average Score')\n",
    "plt.plot(df['DDPG_avg'],'b',label='DDPG Average Score')\n",
    "plt.plot(df['PPO_avg'],'k--' ,label='PPO Average Score')\n",
    "plt.ylim([-650, 100])\n",
    "# plt.plot(df['Averagescores'])\n",
    "plt.ylabel('Average Scores')\n",
    "plt.xlabel('Episode')\n",
    "plt.legend()\n",
    "# plt.xlabel(f'Episode {len(df['DQN_avg'])}')\n",
    "\n",
    "plt.savefig((p_model/'Test_average2.png'))\n",
    "plt.show()\n",
    "df.to_csv((p_model/'scores.csv'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkQAAAHHCAYAAABeLEexAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAABKJklEQVR4nO3deVxVdeL/8fcFWQQFRGQrBEVDcdfMMTdccanUbEwzpVIrU3PLcZwWlaYsLdOZbHFyKVOzJrNyyRBFSxk1HTSVKB23CrBSwCVZz++PfpxvN3ABhQuc1/PxOA85n8/nnPP5nHu9vDnbtRmGYQgAAMDCnBzdAQAAAEcjEAEAAMsjEAEAAMsjEAEAAMsjEAEAAMsjEAEAAMsjEAEAAMsjEAEAAMsjEAEAAMsjEAHA76Snp+uee+5R7dq1ZbPZNH/+fEd3SQkJCbLZbEpISHB0V4Aqi0AEWMzRo0f1yCOPqH79+nJ3d5eXl5c6dOigBQsW6Ndff3V09yRJr732mpYtW+aQbU+aNEmbNm3S9OnTtXz5cvXu3fuqy2RkZMjd3V02m03Jycml3rYjxw1YnY3vMgOsY/369frzn/8sNzc3jRgxQk2bNlVOTo6+/PJLffjhh3rggQe0aNEiR3dTTZs2lZ+fn0OOiAQGBqpHjx569913r3mZf/3rX3r88cfl4+OjkSNH6u9//3uptn25cRcUFCgnJ0eurq5ycuLvWKAsVHN0BwCUj2PHjmnIkCEKDQ3Vli1bFBQUZNaNHTtWR44c0fr16x3Yw4rh9OnT8vHxKdEy7777rvr27avQ0FCtXLmy1IHocpycnOTu7n5D1wngDwwAlvDoo48akowdO3ZcU/vc3FwjNjbWqF+/vuHq6mqEhoYa06dPNy5dumTXTpIxY8aMIsuHhoYaMTEx5vzSpUsNScaXX35pTJo0yfDz8zM8PDyMAQMGGKdPn7ZbTpLd1KVLF8MwDCMnJ8eYOXOm0aBBA8PNzc3w9fU1OnToYHz++edXHc/Ro0eNe+65x6hVq5ZRvXp1o127dsa6deuK9O+P09WcOHHCsNlsxvvvv2/s2rXrivt4+fLlRtu2bY3q1asbPj4+RqdOnYxNmzZdddxbt241JBlbt261W9/7779vtG7d2nB3dzdq165tDBs2zPj+++/t2sTExBienp7G999/b/Tv39/w9PQ0/Pz8jClTphh5eXl2bVetWmW0bt3aqFGjhlGzZk2jadOmxvz586+6D4CqgGOvgEV8+umnql+/vm6//fZraj9q1Cg988wzat26tV555RV16dJFs2fP1pAhQ66rH+PHj9f+/fs1Y8YMjRkzRp9++qnGjRtn1s+fP18333yzGjVqpOXLl2v58uV68sknJUkzZ87UrFmz1LVrV7366qt68sknVbduXe3bt++K20xPT9ftt9+uTZs26bHHHtNzzz2nS5cu6a677tJHH30kSercubOWL18uSerZs6e57atZtWqVPD09dccdd+i2225TeHi4VqxYUaTdrFmzNHz4cLm4uCg2NlazZs1SSEiItmzZctVxF2fZsmUaPHiwnJ2dNXv2bI0ePVpr1qxRx44dlZGRYdc2Pz9f0dHRql27tl566SV16dJFL7/8st3p0bi4OA0dOlS1atXSiy++qBdeeEFRUVHasWPHVfcBUCU4OpEBKHuZmZmGJKN///7X1D4pKcmQZIwaNcqu/IknnjAkGVu2bDHLVMIjRD169DAKCgrM8kmTJhnOzs5GRkaGWdakSRPz6MjvtWjRwujXr981jeH3Jk6caEgyvvjiC7Ps3LlzRr169YywsDAjPz/fbjxjx4695nU3a9bMGDZsmDn/t7/9zfDz8zNyc3PNsu+++85wcnIyBg4caLctwzDs9sXlxv3HI0Q5OTmGv7+/0bRpU+PXX381261bt86QZDzzzDNmWUxMjCHJiI2NtVtnq1atjDZt2pjzEyZMMLy8vIocNQKsgiNEgAVkZWVJkmrWrHlN7Tds2CBJmjx5sl35lClTJOm6rjV6+OGHZbPZzPlOnTopPz9fJ06cuOqyPj4+OnTokL777rsSbXPDhg267bbb1LFjR7OsRo0aevjhh3X8+HEdPny4ROsrdODAAX399dcaOnSoWTZ06FD9/PPP2rRpk1m2du1aFRQU6JlnnilyUfTv98W1+uqrr3T69Gk99thjdtcW9evXT40aNSr29Xn00Uft5jt16qT//e9/5ryPj48uXLiguLi4EvcHqAoIRIAFeHl5SZLOnTt3Te1PnDghJycnNWjQwK48MDBQPj4+1xReLqdu3bp287Vq1ZIknT179qrLxsbGKiMjQ7fccouaNWumqVOn6sCBA1dd7sSJE4qIiChS3rhxY7O+NN599115enqqfv36OnLkiI4cOSJ3d3eFhYXZnTY7evSonJycFBkZWart/FFhf4sbU6NGjYqMx93dXXXq1LErq1Wrlt0+f+yxx3TLLbeoT58+uvnmm/XQQw/ps88+uyH9BSoDAhFgAV5eXgoODtbBgwdLtFxpjl4Uys/PL7bc2dm52HLjGp4A0rlzZx09elRLlixR06ZN9dZbb6l169Z66623St3P0jIMQ6tWrdKFCxcUGRmphg0bmtPx48f18ccf6/z58+Xer+Jcbp//nr+/v5KSkvTJJ5/orrvu0tatW9WnTx/FxMSUQw8BxyMQARZxxx136OjRo0pMTLxq29DQUBUUFBQ5NZWenq6MjAyFhoaaZbVq1SpyEW9OTo5SU1NL3dcrBTFfX189+OCDWrVqlU6dOqXmzZtr5syZV1xfaGioUlJSipR/8803Zn1Jbdu2Td9//71iY2P1wQcf2E2LFi3SxYsXtXbtWklSeHi4CgoKrnpq7loDaGF/ixtTSkpKqcYjSa6urrrzzjv12muvmQ/wfOedd3TkyJFSrQ+oTAhEgEX85S9/kaenp0aNGqX09PQi9UePHtWCBQskSX379pWkIl9bMW/ePEm/XatSKDw8XNu3b7drt2jRosseIboWnp6eRUKWJP3yyy928zVq1FCDBg2UnZ19xfX17dtXu3fvtguDFy5c0KJFixQWFlaqU1mFp8umTp2qe+65x24aPXq0GjZsaJ42GzBggJycnBQbG6uCggK79fz+yNjlxv1Ht956q/z9/fXGG2/YjX3jxo1KTk62e32u1R/3rZOTk5o3by5JV92/QFXAgxkBiwgPD9fKlSt17733qnHjxnZPqt65c6c++OADPfDAA5KkFi1aKCYmRosWLVJGRoa6dOmi3bt36+2339aAAQPUtWtXc72jRo3So48+qkGDBqlnz57av3+/Nm3aJD8/v1L3tU2bNnr99df197//XQ0aNJC/v7+6deumyMhIRUVFqU2bNvL19dVXX32lf//733a37Rfnr3/9q1atWqU+ffro8ccfl6+vr95++20dO3ZMH374YYmf/pydna0PP/xQPXv2vOwDE++66y4tWLBAp0+fVoMGDfTkk0/q2WefVadOnXT33XfLzc1Ne/bsUXBwsGbPnn3Fcf+Ri4uLXnzxRT344IPq0qWLhg4dqvT0dC1YsEBhYWGaNGlSicYj/fY6njlzRt26ddPNN9+sEydO6J///KdatmxpXmsFVGkOvssNQDn79ttvjdGjRxthYWGGq6urUbNmTaNDhw7GP//5T7uHLubm5hqzZs0y6tWrZ7i4uBghISHFPpgxPz/fmDZtmvmgxejoaOPIkSOXve1+z549dssX99DBtLQ0o1+/fkbNmjXtHlD497//3bjtttsMHx8fo3r16kajRo2M5557zsjJybnquAsfzOjj42O4u7sbt912m92DGQvpGm67//DDDw1JxuLFiy/bJiEhwZBkLFiwwCxbsmSJ0apVK8PNzc2oVauW0aVLFyMuLu6q477cgxlXr15trs/X1/eKD2b8oxkzZtg9ePLf//630atXL8Pf399wdXU16tatazzyyCNGamrqFfcFUFXwXWYAAMDyuIYIAABYHoEIAABYHoEIAABYHoEIAABYHoEIAABYHoEIAABYHg9mvAYFBQX68ccfVbNmzev6bicAAFB+DMPQuXPnFBwcfPUHsDryIUjPP/+8ceuttxo1atQw6tSpY/Tv39/45ptv7Nr8+uuvxmOPPWb4+voanp6ext13322kpaXZtTlx4oTRt29fo3r16kadOnWMJ554wsjNzbVrs3XrVqNVq1aGq6urER4ebixduvSa+3nq1ClDEhMTExMTE1MlnE6dOnXV3/UOPUK0bds2jR07Vm3btlVeXp7+9re/qVevXjp8+LA8PT0lSZMmTdL69ev1wQcfyNvbW+PGjdPdd9+tHTt2SPrtG7X79eunwMBA7dy5U6mpqRoxYoRcXFz0/PPPS5KOHTumfv366dFHH9WKFSsUHx+vUaNGKSgoSNHR0VftZ82aNSVJp06dkpeXVxntDQAAcCNlZWUpJCTE/D1+JRXqSdU//fST/P39tW3bNnXu3FmZmZmqU6eOVq5cqXvuuUfSb99O3bhxYyUmJupPf/qTNm7cqDvuuEM//vijAgICJElvvPGGpk2bpp9++kmurq6aNm2a1q9fr4MHD5rbGjJkiDIyMvTZZ59dtV9ZWVny9vZWZmYmgQgAgEqiJL+/K9RF1ZmZmZIkX19fSdLevXuVm5urHj16mG0aNWqkunXrmt9anZiYqGbNmplhSJKio6OVlZWlQ4cOmW1+v47CNr//5uvfy87OVlZWlt0EAACqrgoTiAoKCjRx4kR16NBBTZs2lSSlpaXJ1dVVPj4+dm0DAgKUlpZmtvl9GCqsL6y7UpusrCz9+uuvRfoye/ZseXt7m1NISMgNGSMAAKiYKkwgGjt2rA4ePKj33nvP0V3R9OnTlZmZaU6nTp1ydJcAAEAZqhC33Y8bN07r1q3T9u3bdfPNN5vlgYGBysnJUUZGht1RovT0dAUGBpptdu/ebbe+9PR0s67w38Ky37fx8vJS9erVi/THzc1Nbm5uN2RsAAA4Qn5+vnJzcx3djTLn6up69Vvqr4FDA5FhGBo/frw++ugjJSQkqF69enb1bdq0kYuLi+Lj4zVo0CBJUkpKik6ePKn27dtLktq3b6/nnntOp0+flr+/vyQpLi5OXl5eioyMNNts2LDBbt1xcXHmOgAAqCoMw1BaWpoyMjIc3ZVy4eTkpHr16snV1fW61uPQu8wee+wxrVy5Uh9//LEiIiLMcm9vb/PIzZgxY7RhwwYtW7ZMXl5eGj9+vCRp586dkn5LwC1btlRwcLDmzJmjtLQ0DR8+XKNGjbK77b5p06YaO3asHnroIW3ZskWPP/641q9ff0233XOXGQCgskhNTVVGRob8/f3l4eFRpR8oXPjgZBcXF9WtW7fIWEv0+/uan05YBnSZByj9/qGJhQ9mrFWrluHh4WEMHDjQSE1NtVvP8ePHjT59+hjVq1c3/Pz8jClTphT7YMaWLVsarq6uRv369Uv0YMbMzExDkpGZmXk9wwUAoEzl5eUZhw8fNn7++WdHd6XcZGRkGIcPHzZycnKK1JXk93eFeg5RRcURIgBAZXDp0iUdO3ZMYWFhxV4jWxX9+uuvOn78uOrVqyd3d3e7ukr7HCIAAHD9qvJpsj+6UWMlEAEAAMsjEAEAAIeLiorSxIkTHbb9CvEcIgAAULbK8yxaZbw6mSNEAADA8ghEAADA8ghEAADA8riGCJbiyFtReeQXAFRcHCECAACWRyACAACWRyACAACWRyACAACWRyACAACWx11mAABYQEW/0TUhIcGh2+cIEQAAsDwCEQAAsDwCEQAAsDwCEQAAsDwCEQAAsDwCEQAAsDwCEQAAsDwCEQAAsDwCEQAAsDwCEQAAsDy+ugMAAAtISLCV27aioir494QUgyNEAADA8ghEAADA4bKzs/X444/L399f7u7u6tixo/bs2VNu2ycQAQAAh/vLX/6iDz/8UG+//bb27dunBg0aKDo6WmfOnCmX7ROIAACAQ124cEGvv/665s6dqz59+igyMlL/+te/VL16dS1evLhc+kAgAgAADnX06FHl5uaqQ4cOZpmLi4tuu+02JScnl0sfCEQAAMDyCEQAAMChwsPD5erqqh07dphlubm52rNnjyIjI8ulDzyHCAAAOJSnp6fGjBmjqVOnytfXV3Xr1tWcOXN08eJFjRw5slz6QCACAAAO98ILL6igoEDDhw/XuXPndOutt2rTpk2qVatW+XTAcKBt27YZd9xxhxEUFGRIMj766CO7eknFTnPmzDHbhIaGFqmfPXu23Xr2799vdOzY0XBzczNuvvlm48UXXyxRPzMzMw1JRmZmZqnHiorhcu+p8pgAoKz9+uuvxuHDh41ff/3V0V0pN1cac0l+fzv0GqILFy6oRYsWWrhwYbH1qampdtOSJUtks9k0aNAgu3axsbF27caPH2/WZWVlqVevXgoNDdXevXs1d+5czZw5U4sWLSrTsQEAgMrDoafM+vTpoz59+ly2PjAw0G7+448/VteuXVW/fn278po1axZpW2jFihXKycnRkiVL5OrqqiZNmigpKUnz5s3Tww8/fP2DAAAAlV6lucssPT1d69evL/biqhdeeEG1a9dWq1atNHfuXOXl5Zl1iYmJ6ty5s1xdXc2y6OhopaSk6OzZs+XSdwAAULFVmouq3377bdWsWVN33323Xfnjjz+u1q1by9fXVzt37tT06dOVmpqqefPmSZLS0tJUr149u2UCAgLMuuIu1srOzlZ2drY5n5WVdaOHAwAAKpBKE4iWLFmiYcOGyd3d3a588uTJ5s/NmzeXq6urHnnkEc2ePVtubm6l2tbs2bM1a9as6+ovAACoPCrFKbMvvvhCKSkpGjVq1FXbtmvXTnl5eTp+/Lik365DSk9Pt2tTOH+5646mT5+uzMxMczp16tT1DQAAAFRolSIQLV68WG3atFGLFi2u2jYpKUlOTk7y9/eXJLVv317bt29Xbm6u2SYuLk4RERGXfbaBm5ubvLy87CYAAFB1OTQQnT9/XklJSUpKSpIkHTt2TElJSTp58qTZJisrSx988EGxR4cSExM1f/587d+/X//73/+0YsUKTZo0Sffff78Zdu677z65urpq5MiROnTokFavXq0FCxbYnWoDAADW5tBriL766it17drVnC8MKTExMVq2bJkk6b333pNhGBo6dGiR5d3c3PTee+9p5syZys7OVr169TRp0iS7sOPt7a3PP/9cY8eOVZs2beTn56dnnnmGW+4BAIDJZhiG4ehOVHRZWVny9vZWZmYmp88qOZvN5rBt818NQFm7dOmSjh07pnr16hW5CamqutKYS/L7u9LcZQYAAEqvPP8grIx/AFaKi6oBAADKEoEIAAA41Lp16+Tj46P8/HxJv90xbrPZ9Ne//tVsM2rUKN1///1l1gcCEQAAcKhOnTrp3Llz+u9//ytJ2rZtm/z8/JSQkGC22bZtm6KiosqsDwQiAADgUN7e3mrZsqUZgBISEjRp0iT997//1fnz5/XDDz/oyJEj6tKlS5n1gUAEAAAcrkuXLkpISJBhGPriiy909913q3Hjxvryyy+1bds2BQcHq2HDhmW2fe4yAwAADhcVFaUlS5Zo//79cnFxUaNGjRQVFaWEhASdPXu2TI8OSRwhAgAAFUDhdUSvvPKKGX4KA1FCQkKZXj8kEYgAAEAFUKtWLTVv3lwrVqwww0/nzp21b98+ffvttxwhAgAA1tClSxfl5+ebgcjX11eRkZEKDAxUREREmW6bQAQAgAUYhlFuU2nNnz9fhmGoUaNGZllSUpJSU1NvxC64IgIRAACwPAIRAACwPAIRAACwPAIRAACwPAIRAABVzPVc2FzZ3KixEogAAKgiXFxcJEkXL150cE/KT05OjiTJ2dn5utbDV3cAAFBFODs7y8fHR6dPn5YkeXh4yGazObhXZaegoEA//fSTPDw8VK3a9UUaAhEAAFVIYGCgJJmhqKpzcnJS3bp1rzv4EYgAAKhCbDabgoKC5O/vr9zcXEd3p8y5urrKyen6rwAiEAEAUAU5Oztf93U1VsJF1QAAwPIIRAAAwPIIRAAAwPIIRAAAwPIIRAAAwPIIRAAAwPIIRAAAwPIIRAAAwPIIRAAAwPIIRAAAwPIIRAAAwPIIRAAAwPIIRAAAwPIIRAAAwPIcGoi2b9+uO++8U8HBwbLZbFq7dq1d/QMPPCCbzWY39e7d267NmTNnNGzYMHl5ecnHx0cjR47U+fPn7docOHBAnTp1kru7u0JCQjRnzpyyHhoAAKhEHBqILly4oBYtWmjhwoWXbdO7d2+lpqaa06pVq+zqhw0bpkOHDikuLk7r1q3T9u3b9fDDD5v1WVlZ6tWrl0JDQ7V3717NnTtXM2fO1KJFi8psXAAAoHKp5siN9+nTR3369LliGzc3NwUGBhZbl5ycrM8++0x79uzRrbfeKkn65z//qb59++qll15ScHCwVqxYoZycHC1ZskSurq5q0qSJkpKSNG/ePLvgBAAArKvCX0OUkJAgf39/RUREaMyYMfrll1/MusTERPn4+JhhSJJ69OghJycn7dq1y2zTuXNnubq6mm2io6OVkpKis2fPFrvN7OxsZWVl2U0AAKDqqtCBqHfv3nrnnXcUHx+vF198Udu2bVOfPn2Un58vSUpLS5O/v7/dMtWqVZOvr6/S0tLMNgEBAXZtCucL2/zR7Nmz5e3tbU4hISE3emgAAKACcegps6sZMmSI+XOzZs3UvHlzhYeHKyEhQd27dy+z7U6fPl2TJ08257OysghFAABUYRX6CNEf1a9fX35+fjpy5IgkKTAwUKdPn7Zrk5eXpzNnzpjXHQUGBio9Pd2uTeH85a5NcnNzk5eXl90EAACqrkoViL7//nv98ssvCgoKkiS1b99eGRkZ2rt3r9lmy5YtKigoULt27cw227dvV25urtkmLi5OERERqlWrVvkOAAAAVEgODUTnz59XUlKSkpKSJEnHjh1TUlKSTp48qfPnz2vq1Kn6z3/+o+PHjys+Pl79+/dXgwYNFB0dLUlq3LixevfurdGjR2v37t3asWOHxo0bpyFDhig4OFiSdN9998nV1VUjR47UoUOHtHr1ai1YsMDulBgAALA4w4G2bt1qSCoyxcTEGBcvXjR69epl1KlTx3BxcTFCQ0ON0aNHG2lpaXbr+OWXX4yhQ4caNWrUMLy8vIwHH3zQOHfunF2b/fv3Gx07djTc3NyMm266yXjhhRdK1M/MzExDkpGZmXndY4ZjFfd+K68JAFC+SvL722YYhlGWgasqyMrKkre3tzIzM7meqJKz2WwO2zb/1QCgfJXk93eluoYIAACgLBCIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5RGIAACA5Tk0EG3fvl133nmngoODZbPZtHbtWrMuNzdX06ZNU7NmzeTp6ang4GCNGDFCP/74o906wsLCZLPZ7KYXXnjBrs2BAwfUqVMnubu7KyQkRHPmzCmP4QEAgErCoYHowoULatGihRYuXFik7uLFi9q3b5+efvpp7du3T2vWrFFKSoruuuuuIm1jY2OVmppqTuPHjzfrsrKy1KtXL4WGhmrv3r2aO3euZs6cqUWLFpXp2AAAQOVRzZEb79Onj/r06VNsnbe3t+Li4uzKXn31Vd122206efKk6tata5bXrFlTgYGBxa5nxYoVysnJ0ZIlS+Tq6qomTZooKSlJ8+bN08MPP3zjBgMAACqtSnUNUWZmpmw2m3x8fOzKX3jhBdWuXVutWrXS3LlzlZeXZ9YlJiaqc+fOcnV1Ncuio6OVkpKis2fPFrud7OxsZWVl2U0AAKDqcugRopK4dOmSpk2bpqFDh8rLy8ssf/zxx9W6dWv5+vpq586dmj59ulJTUzVv3jxJUlpamurVq2e3roCAALOuVq1aRbY1e/ZszZo1qwxHAwAAKpJKEYhyc3M1ePBgGYah119/3a5u8uTJ5s/NmzeXq6urHnnkEc2ePVtubm6l2t706dPt1puVlaWQkJDSdR4AAFR4FT4QFYahEydOaMuWLXZHh4rTrl075eXl6fjx44qIiFBgYKDS09Pt2hTOX+66Izc3t1KHKQAAUPlU6GuICsPQd999p82bN6t27dpXXSYpKUlOTk7y9/eXJLVv317bt29Xbm6u2SYuLk4RERHFni4DAADW49AjROfPn9eRI0fM+WPHjikpKUm+vr4KCgrSPffco3379mndunXKz89XWlqaJMnX11eurq5KTEzUrl271LVrV9WsWVOJiYmaNGmS7r//fjPs3HfffZo1a5ZGjhypadOm6eDBg1qwYIFeeeUVh4wZAABUPDbDMAxHbTwhIUFdu3YtUh4TE6OZM2cWuRi60NatWxUVFaV9+/bpscce0zfffKPs7GzVq1dPw4cP1+TJk+1OeR04cEBjx47Vnj175Ofnp/Hjx2vatGnX3M+srCx5e3srMzPzqqfsULHZbDaHbduB/9UAwJJK8vvboYGosiAQVR0EIgCwjpL8/q7Q1xABAACUBwIRAACwPAIRAACwPAIRAACwPAIRAACwPAIRAACwPAIRAACwPAIRAACwPAIRAACwPAIRAACwPAIRAACwPAIRAACwPAIRAACwPAIRAACwPAIRAACwPAIRAACwPAIRAACwPAIRAACwPAIRAACwvFIFovr16+uXX34pUp6RkaH69etfd6cAAADKU6kC0fHjx5Wfn1+kPDs7Wz/88MN1dwoAAKA8VStJ408++cT8edOmTfL29jbn8/PzFR8fr7CwsBvWOQAAgPJQokA0YMAASZLNZlNMTIxdnYuLi8LCwvTyyy/fsM4BAACUhxIFooKCAklSvXr1tGfPHvn5+ZVJpwAAAMpTiQJRoWPHjt3ofgAAADhMqQKRJMXHxys+Pl6nT582jxwVWrJkyXV3DAAAoLyUKhDNmjVLsbGxuvXWWxUUFCSbzXaj+wUAAFBuShWI3njjDS1btkzDhw+/0f0BAAAod6V6DlFOTo5uv/32G90XAAAAhyhVIBo1apRWrlx5o/sCAADgEKU6ZXbp0iUtWrRImzdvVvPmzeXi4mJXP2/evBvSOQAAgPJQqkB04MABtWzZUpJ08OBBuzousAYAAJVNqQLR1q1bb3Q/AAAAHKZU1xABAABUJaUKRF27dlW3bt0uO12r7du3684771RwcLBsNpvWrl1rV28Yhp555hkFBQWpevXq6tGjh7777ju7NmfOnNGwYcPk5eUlHx8fjRw5UufPn7drc+DAAXXq1Enu7u4KCQnRnDlzSjNsAABQRZUqELVs2VItWrQwp8jISOXk5Gjfvn1q1qzZNa/nwoULatGihRYuXFhs/Zw5c/SPf/xDb7zxhnbt2iVPT09FR0fr0qVLZpthw4bp0KFDiouL07p167R9+3Y9/PDDZn1WVpZ69eql0NBQ7d27V3PnztXMmTO1aNGi0gwdAABURcYNNGPGDGPKlCmlWlaS8dFHH5nzBQUFRmBgoDF37lyzLCMjw3BzczNWrVplGIZhHD582JBk7Nmzx2yzceNGw2azGT/88INhGIbx2muvGbVq1TKys7PNNtOmTTMiIiKuuW+ZmZmGJCMzM7NUY0PFIclhEwCgfJXk9/cNvYbo/vvvv2HfY3bs2DGlpaWpR48eZpm3t7fatWunxMRESVJiYqJ8fHx06623mm169OghJycn7dq1y2zTuXNnubq6mm2io6OVkpKis2fP3pC+AgCAyq3UX+5anMTERLm7u9+QdaWlpUmSAgIC7MoDAgLMurS0NPn7+9vVV6tWTb6+vnZt6tWrV2QdhXW1atUqsu3s7GxlZ2eb81lZWdc5GgAAUJGVKhDdfffddvOGYSg1NVVfffWVnn766RvSMUeaPXu2Zs2a5ehuAACAclKqU2be3t52k6+vr6KiorRhwwbNmDHjhnQsMDBQkpSenm5Xnp6ebtYFBgbq9OnTdvV5eXk6c+aMXZvi1vH7bfzR9OnTlZmZaU6nTp26/gEBAIAKq1RHiJYuXXqj+1FEvXr1FBgYqPj4ePOp2FlZWdq1a5fGjBkjSWrfvr0yMjK0d+9etWnTRpK0ZcsWFRQUqF27dmabJ598Urm5ueZXjMTFxSkiIqLY02WS5ObmJjc3tzIeIQAAqCiu6xqivXv3Kjk5WZLUpEkTtWrVqkTLnz9/XkeOHDHnjx07pqSkJPn6+qpu3bqaOHGi/v73v6thw4aqV6+enn76aQUHB2vAgAGSpMaNG6t3794aPXq03njjDeXm5mrcuHEaMmSIgoODJUn33XefZs2apZEjR2ratGk6ePCgFixYoFdeeeV6hg4AAKqS0tzGlp6ebnTt2tWw2WxGrVq1jFq1ahk2m83o1q2bcfr06Wtez9atW4u9PTkmJsYwjN9uvX/66aeNgIAAw83NzejevbuRkpJit45ffvnFGDp0qFGjRg3Dy8vLePDBB41z587Ztdm/f7/RsWNHw83NzbjpppuMF154oUTj5bb7qqO491t5TQCA8lWS3982wzCMkoaoe++9V//73//0zjvvqHHjxpKkw4cPKyYmRg0aNNCqVatKusoKLSsrS97e3srMzJSXl5eju4Pr4MgvHy7FfzUAwHUoye/vUgUib29vbd68WW3btrUr3717t3r16qWMjIySrrJCIxBVHQQiALCOkvz+LtVdZgUFBeYFyr/n4uKigoKC0qwSAADAYUoViLp166YJEyboxx9/NMt++OEHTZo0Sd27d79hnQMAACgPpQpEr776qrKyshQWFqbw8HCFh4erXr16ysrK0j//+c8b3UcAAIAyVarb7kNCQrRv3z5t3rxZ33zzjaTfboH//feOAQAAVBYlOkK0ZcsWRUZGKisrSzabTT179tT48eM1fvx4tW3bVk2aNNEXX3xRVn0FAAAoEyUKRPPnz9fo0aOLvVLb29tbjzzyiObNm3fDOgcAAFAeShSI9u/fr969e1+2vlevXtq7d+91dwoAAKA8lSgQpaenF3u7faFq1arpp59+uu5OAQAAlKcSBaKbbrpJBw8evGz9gQMHFBQUdN2dAgAAKE8lCkR9+/bV008/rUuXLhWp+/XXXzVjxgzdcccdN6xzAAAA5aFEX92Rnp6u1q1by9nZWePGjVNERIQk6ZtvvtHChQuVn5+vffv2KSAgoMw67Ah8dUfVwVd3AIB1lOT3d4meQxQQEKCdO3dqzJgxmj59uvkBb7PZFB0drYULF1a5MAQAAKq+Ej+YMTQ0VBs2bNDZs2d15MgRGYahhg0bqlatWmXRPwAAgDJXqidVS1KtWrWKfNs9AABAZVSq7zIDAACoSghEAADA8ghEAADA8ghEAADA8ghEAADA8ghEAADA8ghEAADA8ghEAADA8ghEAADA8ghEAADA8ghEAADA8ghEAADA8ghEAADA8ghEAADA8ghEAADA8ghEAADA8ghEAADA8ghEAADA8ghEAADA8ghEAADA8ip8IAoLC5PNZisyjR07VpIUFRVVpO7RRx+1W8fJkyfVr18/eXh4yN/fX1OnTlVeXp4jhgMAACqgao7uwNXs2bNH+fn55vzBgwfVs2dP/fnPfzbLRo8erdjYWHPew8PD/Dk/P1/9+vVTYGCgdu7cqdTUVI0YMUIuLi56/vnny2cQAACgQqvwgahOnTp28y+88ILCw8PVpUsXs8zDw0OBgYHFLv/555/r8OHD2rx5swICAtSyZUs9++yzmjZtmmbOnClXV9cy7T8AAKj4Kvwps9/LycnRu+++q4ceekg2m80sX7Fihfz8/NS0aVNNnz5dFy9eNOsSExPVrFkzBQQEmGXR0dHKysrSoUOHyrX/AACgYqrwR4h+b+3atcrIyNADDzxglt13330KDQ1VcHCwDhw4oGnTpiklJUVr1qyRJKWlpdmFIUnmfFpaWrHbyc7OVnZ2tjmflZV1g0cCAAAqkkoViBYvXqw+ffooODjYLHv44YfNn5s1a6agoCB1795dR48eVXh4eKm2M3v2bM2aNeu6+wsAACqHSnPK7MSJE9q8ebNGjRp1xXbt2rWTJB05ckSSFBgYqPT0dLs2hfOXu+5o+vTpyszMNKdTp05db/cBAEAFVmkC0dKlS+Xv769+/fpdsV1SUpIkKSgoSJLUvn17ff311zp9+rTZJi4uTl5eXoqMjCx2HW5ubvLy8rKbAABA1VUpTpkVFBRo6dKliomJUbVq/9flo0ePauXKlerbt69q166tAwcOaNKkSercubOaN28uSerVq5ciIyM1fPhwzZkzR2lpaXrqqac0duxYubm5OWpIAACgAqkUgWjz5s06efKkHnroIbtyV1dXbd68WfPnz9eFCxcUEhKiQYMG6amnnjLbODs7a926dRozZozat28vT09PxcTE2D23CAAAWJvNMAzD0Z2o6LKysuTt7a3MzExOn1Vyv39cQ3njvxoAlK+S/P6uNNcQAQAAlBUCEQAAsDwCEQAAsDwCEQAAsDwCEQAAsDwCEQAAsDwCEQAAsDwCEQAAsDwCEQAAsLxK8dUdAK5NQoLjnsQtSVFRPI0bN54j39e8p62DI0QAAMDyCEQAAMDyCEQAAMDyCEQAAMDyCEQAAMDyCEQAAMDyCEQAAMDyCEQAAMDyCEQAAMDyCEQAAMDyCEQAAMDyCEQAAMDyCEQAAMDyCEQAAMDyqjm6AwBQGSUk2By6/agow6HbB6oajhABAADL4wgRAACw/FFPjhABAADL4wgRyoUj//Jw9F8dAICKjyNEAADA8ghEAADA8ghEAADA8ghEAADA8ghEAADA8ghEAADA8ip0IJo5c6ZsNpvd1KhRI7P+0qVLGjt2rGrXrq0aNWpo0KBBSk9Pt1vHyZMn1a9fP3l4eMjf319Tp05VXl5eeQ8FAABUYBX+OURNmjTR5s2bzflq1f6vy5MmTdL69ev1wQcfyNvbW+PGjdPdd9+tHTt2SJLy8/PVr18/BQYGaufOnUpNTdWIESPk4uKi559/vtzHAgAAKqYKH4iqVaumwMDAIuWZmZlavHixVq5cqW7dukmSli5dqsaNG+s///mP/vSnP+nzzz/X4cOHtXnzZgUEBKhly5Z69tlnNW3aNM2cOVOurq7lPRwAAFABVehTZpL03XffKTg4WPXr19ewYcN08uRJSdLevXuVm5urHj16mG0bNWqkunXrKjExUZKUmJioZs2aKSAgwGwTHR2trKwsHTp06LLbzM7OVlZWlt0EAACqrgodiNq1a6dly5bps88+0+uvv65jx46pU6dOOnfunNLS0uTq6iofHx+7ZQICApSWliZJSktLswtDhfWFdZcze/ZseXt7m1NISMiNHRgAAKhQKvQpsz59+pg/N2/eXO3atVNoaKjef/99Va9evcy2O336dE2ePNmcz8rKIhQBAFCFVegjRH/k4+OjW265RUeOHFFgYKBycnKUkZFh1yY9Pd285igwMLDIXWeF88Vdl1TIzc1NXl5edhMAAKi6KlUgOn/+vI4ePaqgoCC1adNGLi4uio+PN+tTUlJ08uRJtW/fXpLUvn17ff311zp9+rTZJi4uTl5eXoqMjCz3/gMAgIqpQp8ye+KJJ3TnnXcqNDRUP/74o2bMmCFnZ2cNHTpU3t7eGjlypCZPnixfX195eXlp/Pjxat++vf70pz9Jknr16qXIyEgNHz5cc+bMUVpamp566imNHTtWbm5uDh4dAACoKCp0IPr+++81dOhQ/fLLL6pTp446duyo//znP6pTp44k6ZVXXpGTk5MGDRqk7OxsRUdH67XXXjOXd3Z21rp16zRmzBi1b99enp6eiomJUWxsrKOGBAAAKqAKHYjee++9K9a7u7tr4cKFWrhw4WXbhIaGasOGDTe6awAAoAqpVNcQAQAAlAUCEQAAsDwCEQAAsDwCEQAAsDwCEQAAsDwCEQAAsDwCEQAAsDwCEQAAsDwCEQAAsDwCEQAAsDwCEQDgimw2x05AeSAQAQAAyyMQAQAAyyMQAQAAyyMQAQAAyyMQAQAAyyMQAQAAyyMQAQAAyyMQAQAAyyMQAQAAyyMQAQAAyyMQAQAAyyMQAQAAyyMQAQAAyyMQAQAAyyMQAQAAyyMQAQAAyyMQAQAAyyMQAQAAyyMQAQAAyyMQAQAAyyMQAQAAyyMQAQAAyyMQAQAAyyMQAQAAy6vQgWj27Nlq27atatasKX9/fw0YMEApKSl2baKiomSz2eymRx991K7NyZMn1a9fP3l4eMjf319Tp05VXl5eeQ4FAABUYNUc3YEr2bZtm8aOHau2bdsqLy9Pf/vb39SrVy8dPnxYnp6eZrvRo0crNjbWnPfw8DB/zs/PV79+/RQYGKidO3cqNTVVI0aMkIuLi55//vlyHQ8AAKiYKnQg+uyzz+zmly1bJn9/f+3du1edO3c2yz08PBQYGFjsOj7//HMdPnxYmzdvVkBAgFq2bKlnn31W06ZN08yZM+Xq6lqmYwAAABVfhT5l9keZmZmSJF9fX7vyFStWyM/PT02bNtX06dN18eJFsy4xMVHNmjVTQECAWRYdHa2srCwdOnSo2O1kZ2crKyvLbgIAAFVXhT5C9HsFBQWaOHGiOnTooKZNm5rl9913n0JDQxUcHKwDBw5o2rRpSklJ0Zo1ayRJaWlpdmFIkjmflpZW7LZmz56tWbNmldFIAABARVNpAtHYsWN18OBBffnll3blDz/8sPlzs2bNFBQUpO7du+vo0aMKDw8v1bamT5+uyZMnm/NZWVkKCQkpXccBAECFVylOmY0bN07r1q3T1q1bdfPNN1+xbbt27SRJR44ckSQFBgYqPT3drk3h/OWuO3Jzc5OXl5fdBAAAqq4KHYgMw9C4ceP00UcfacuWLapXr95Vl0lKSpIkBQUFSZLat2+vr7/+WqdPnzbbxMXFycvLS5GRkWXSbwAAULlU6FNmY8eO1cqVK/Xxxx+rZs2a5jU/3t7eql69uo4ePaqVK1eqb9++ql27tg4cOKBJkyapc+fOat68uSSpV69eioyM1PDhwzVnzhylpaXpqaee0tixY+Xm5ubI4QEAgAqiQh8hev3115WZmamoqCgFBQWZ0+rVqyVJrq6u2rx5s3r16qVGjRppypQpGjRokD799FNzHc7Ozlq3bp2cnZ3Vvn173X///RoxYoTdc4sAAIC1VegjRIZhXLE+JCRE27Ztu+p6QkNDtWHDhhvVLQAAUMVU6CNEAAAA5YFABAAALI9ABAAALI9ABAAALI9ABKBSstkcOwGoWghEAADA8ghEAADA8ghEAADA8ghEAADA8ghEAADA8ghEAADA8ghEAADA8ghEAADA8ghEAABUADxs1LEIRAAAwPIIRBbBXx4AAFwegQgAAFgegQgAAFgegQgAAFgegQgAAFgegQgAAFgegQgoJ9zNBwAVF4EIAABYHoEIAABYHoEIAABYHoEIAABYHoEIAABYHoEIAABYHoEIAABYHoEIAABYHoEIAABYHoEIAABYHoEIAABYHoEIAABYnqUC0cKFCxUWFiZ3d3e1a9dOu3fvdnSXAABABWCZQLR69WpNnjxZM2bM0L59+9SiRQtFR0fr9OnTju4aAABwMMsEonnz5mn06NF68MEHFRkZqTfeeEMeHh5asmSJo7sGAAAczBKBKCcnR3v37lWPHj3MMicnJ/Xo0UOJiYkO7BkAAKgIqjm6A+Xh559/Vn5+vgICAuzKAwIC9M033xRpn52drezsbHM+MzNTkpSVlVW2Ha3CLlxw3LYrzutW9v1w5H6WKtK+Lnvs6/LD50f5qIrv6cJ1GoZx1baWCEQlNXv2bM2aNatIeUhIiAN6UzXccYcjt+7tyI3/Ttn3w7H7Wao4+7rssa/LD58f5aMqv6fPnTsnb+8rr98SgcjPz0/Ozs5KT0+3K09PT1dgYGCR9tOnT9fkyZPN+YKCAp05c0a1a9eWzWYr8/5WNVlZWQoJCdGpU6fk5eXl6O5UWezn8sO+Lj/s6/JRVfezYRg6d+6cgoODr9rWEoHI1dVVbdq0UXx8vAYMGCDpt5ATHx+vcePGFWnv5uYmNzc3uzIfH59y6GnV5uXlVaX+o1VU7Ofyw74uP+zr8lEV9/PVjgwVskQgkqTJkycrJiZGt956q2677TbNnz9fFy5c0IMPPujorgEAAAezTCC699579dNPP+mZZ55RWlqaWrZsqc8++6zIhdYAAMB6LBOIJGncuHHFniJD2XJzc9OMGTOKnIbEjcV+Lj/s6/LDvi4f7GfJZlzLvWgAAABVmCUezAgAAHAlBCIAAGB5BCIAAGB5BCJUCMuWLbvqs54eeOAB8zlSKDmbzaa1a9eWaJmZM2cqICCgVMtaWUJCgmw2mzIyMq55mZkzZ6ply5Zl1qfKJCoqShMnTnR0NyzPaq8DgaiKeeCBB2Sz2WSz2eTi4qKAgAD17NlTS5YsUUFBgaO7J0kKCwvT/PnzS7zcggULtGzZshven7JU+Fpcbpo5c6aju3hZycnJmjVrlt58802lpqaqT58+pX7tbrSKvl9vv/12paamXvMD4a5Vef+CSkxMlLOzs/r161du2yxvlT3s//4z32azqXbt2urdu7cOHDjg6K5VOgSiKqh3795KTU3V8ePHtXHjRnXt2lUTJkzQHXfcoby8PEd3r9S8vb0r3RPDU1NTzWn+/Pny8vKyK3viiSdKtL6cnJwy6mlRR48elST1799fgYGBFep23Iq8X3Nzc+Xq6qrAwMBK/1U/ixcv1vjx47V9+3b9+OOPju7ONcvPz68wfwCWh8LP/NTUVMXHx6tatWq64wpfTJabm1uOvas8CERVkJubmwIDA3XTTTepdevW+tvf/qaPP/5YGzdutDvCcvLkSfXv3181atSQl5eXBg8eXOT73n7v+PHjstlsWrNmjbp27SoPDw+1aNFCiYmJdu0+/PBDNWnSRG5ubgoLC9PLL79s1kVFRenEiROaNGmS+RfN723atEmNGzdWjRo1zP/khf54yiwqKkqPP/64/vKXv8jX11eBgYFFjgx888036tixo9zd3RUZGanNmzeX61+EgYGB5uTt7S2bzWbOv/HGG+rYsaNd+/nz5yssLMycLxzzc889p+DgYEVERFzz63A1p06d0uDBg+Xj4yNfX1/1799fx48fl/Tb6Zs777xTkuTk5CSbzXbV1648VaT9arPZ9Prrr+uuu+6Sp6ennnvuuWJPmf3rX/9SSEiIPDw8NHDgQM2bN6/YgL98+XKFhYXJ29tbQ4YM0blz58w+b9u2TQsWLDD3f+HrVRbOnz+v1atXa8yYMerXr1+Ro7Offvqp2rZtK3d3d/n5+WngwIFmXXZ2tqZNm6aQkBC5ubmpQYMGWrx4sVl/8OBB9enTRzVq1FBAQICGDx+un3/++bJ9yc7O1hNPPKGbbrpJnp6eateunRISEsz6wlPun3zyiSIjI+Xm5qaTJ09qz5496tmzp/z8/OTt7a0uXbpo37595nKF74mBAwfKZrPZvUc+/vhjtW7dWu7u7qpfv75mzZpVYf+gLPzMDwwMVMuWLfXXv/5Vp06d0k8//WS+r1evXq0uXbrI3d1dK1as0C+//KKhQ4fqpptukoeHh5o1a6ZVq1ZdcTvr16+Xt7e3VqxYIenKnyGVEYHIIrp166YWLVpozZo1kn77Lrf+/fvrzJkz2rZtm+Li4vS///1P995771XX9eSTT+qJJ55QUlKSbrnlFg0dOtT8oNi7d68GDx6sIUOG6Ouvv9bMmTP19NNPmx+ma9as0c0336zY2FjzL5pCFy9e1EsvvaTly5dr+/btOnny5FX/0n/77bfl6empXbt2ac6cOYqNjVVcXJyk3/5KHDBggDw8PLRr1y4tWrRITz75ZGl2n0PFx8crJSVFcXFxWrdunVl+pdfhanJzcxUdHa2aNWvqiy++0I4dO8wQmpOToyeeeEJLly6V9H9HY6702lVGN3K/zpw5UwMHDtTXX3+thx56qEj9jh079Oijj2rChAlKSkpSz5499dxzzxVpd/ToUa1du1br1q3TunXrtG3bNr3wwguSfjtl3L59e40ePdrc/yEhIde5Fy7v/fffV6NGjRQREaH7779fS5YsUeFj69avX6+BAweqb9+++u9//6v4+Hjddttt5rIjRozQqlWr9I9//EPJycl68803VaNGDUlSRkaGunXrplatWumrr77SZ599pvT0dA0ePPiyfRk3bpwSExP13nvv6cCBA/rzn/+s3r1767vvvjPbXLx4US+++KLeeustHTp0SP7+/jp37pxiYmL05Zdf6j//+Y8aNmyovn37miFzz549kqSlS5cqNTXVnP/iiy80YsQITZgwQYcPH9abb76pZcuWFfuaVTTnz5/Xu+++qwYNGqh27dpm+V//+ldNmDBBycnJio6O1qVLl9SmTRutX79eBw8e1MMPP6zhw4dr9+7dxa535cqVGjp0qFasWKFhw4Zd9TOkUjJQpcTExBj9+/cvtu7ee+81GjdubBiGYXz++eeGs7OzcfLkSbP+0KFDhiRj9+7dxS5/7NgxQ5Lx1ltvFVkmOTnZMAzDuO+++4yePXvaLTd16lQjMjLSnA8NDTVeeeUVuzZLly41JBlHjhwxyxYuXGgEBARcdmxdunQxOnbsaLeetm3bGtOmTTMMwzA2btxoVKtWzUhNTTXr4+LiDEnGRx99VOwYy9LSpUsNb29vc37GjBlGixYt7Nq88sorRmhoqDkfExNjBAQEGNnZ2WbZtbwOxfn9uJcvX25EREQYBQUFZn12drZRvXp1Y9OmTYZhGMZHH31k/PEjorjXztEqwn6dOHGiXdnWrVsNScbZs2cNw/jt/16/fv3s2gwbNqxIvz08PIysrCyzbOrUqUa7du3M+S5duhgTJky4bF9upNtvv92YP3++YRiGkZuba/j5+Rlbt241DMMw2rdvbwwbNqzY5VJSUgxJRlxcXLH1zz77rNGrVy+7slOnThmSjJSUFMMw7Md54sQJw9nZ2fjhhx/slunevbsxffp0wzD+7/MjKSnpimPKz883atasaXz66admWXGfB927dzeef/55u7Lly5cbQUFBV1y/I8TExBjOzs6Gp6en4enpaUgygoKCjL179xqG8X/v68LX8kr69etnTJkyxZwvfB1effVVw9vb20hISDDrruUzpLLhCJGFGIZhnuZITk5WSEiI3V+YkZGR8vHxUXJy8hXX07x5c/PnoKAgSdLp06fN9Xbo0MGufYcOHfTdd98pPz//iuv18PBQeHi43boL13stffnjMikpKQoJCVFgYKBZ//u/YiuLZs2aydXVtUj5lV6Hq9m/f7+OHDmimjVrqkaNGqpRo4Z8fX116dIl89qhqu5G7tdbb731ivUpKSlF3nvFvRfDwsJUs2ZNu+1f62t6I6WkpGj37t0aOnSoJKlatWq69957zdNeSUlJ6t69e7HLJiUlydnZWV26dCm2fv/+/dq6dav5vqtRo4YaNWokScW+977++mvl5+frlltusVtm27Ztdu1dXV2LfB6kp6dr9OjRatiwoby9veXl5aXz58/r5MmTVxz//v37FRsba7e9wiNzFy9evOKyjtC1a1clJSUpKSlJu3fvVnR0tPr06aMTJ06Ybf74Hs3Pz9ezzz6rZs2aydfXVzVq1NCmTZuK7Jt///vfmjRpkuLi4uxe06r4GWKp7zKzuuTkZNWrV++61+Pi4mL+XBiwbsQFjL9fb+G6jat8s0xxy1SWiymdnJyKjK+4ix09PT2LXf56Xofz58+rTZs25rUAv1enTp1rWkdF5Yj9erl1lVRFeT8vXrxYeXl5Cg4ONssMw5Cbm5teffVVVa9e/bLLXqlO+u29d+edd+rFF18sUlcYQP/Y3tnZWXv37pWzs7NdXeFpuMLt/vG6tpiYGP3yyy9asGCBQkND5ebmpvbt21/1lM758+c1a9Ys3X333UXq3N3dr7isI3h6eqpBgwbm/FtvvSVvb2/961//0qhRo8w2vzd37lwtWLBA8+fPV7NmzeTp6amJEycW2TetWrXSvn37tGTJEt16663mPq6KnyEEIovYsmWLvv76a02aNEmS1LhxY506dUqnTp0yjxIdPnxYGRkZioyMLPV2GjdurB07dtiV7dixQ7fccov5Yebq6nrVo0U3QkREhE6dOqX09HQFBARI+r9rBiqCOnXqKC0tze7IXVJSUrlsu3Xr1lq9erX8/f3l5eV1zcuV12t3PRy5Xy8nIiKiyHuvNO/F8tj/eXl5euedd/Tyyy+rV69ednUDBgzQqlWr1Lx5c8XHx+vBBx8ssnyzZs1UUFCgbdu2qUePHkXqW7durQ8//FBhYWGqVu3qv4JatWql/Px8nT59Wp06dSrRWHbs2KHXXntNffv2lfTbRcB/vHjbxcWlyD5t3bq1UlJS7EJGZWKz2eTk5KRff/31sm127Nih/v376/7775f0W+j/9ttvi3z+h4eH6+WXX1ZUVJScnZ316quvSir9Z0hFximzKig7O1tpaWn64YcftG/fPj3//PPq37+/7rjjDo0YMUKS1KNHDzVr1kzDhg3Tvn37tHv3bo0YMUJdunS56uH/K5kyZYri4+P17LPP6ttvv9Xbb7+tV1991e7i6LCwMG3fvl0//PDDFe8suV49e/ZUeHi4YmJidODAAe3YsUNPPfWUJFWI26GjoqL0008/ac6cOTp69KgWLlyojRs3lsu2hw0bJj8/P/Xv319ffPGFjh07poSEBD3++OP6/vvvL7tceb1218OR+/Vyxo8frw0bNmjevHn67rvv9Oabb2rjxo0lfh+GhYVp165dOn78uH7++ecyOXq0bt06nT17ViNHjlTTpk3tpkGDBmnx4sWaMWOGVq1apRkzZig5OVlff/21ecQnLCxMMTExeuihh7R27VrzvfX+++9LksaOHaszZ85o6NCh2rNnj44ePapNmzbpwQcfLDbs3XLLLRo2bJhGjBihNWvW6NixY9q9e7dmz56t9evXX3EsDRs21PLly5WcnKxdu3Zp2LBhRY5ghYWFKT4+XmlpaTp79qwk6ZlnntE777yjWbNm6dChQ0pOTtZ7771nfn5UNIWf+WlpaUpOTtb48ePNI3GX07BhQ8XFxWnnzp1KTk7WI488ctm7jG+55RZt3bpVH374ofkcrNJ+hlRkBKIq6LPPPlNQUJDCwsLUu3dvbd26Vf/4xz/08ccfm0dpbDabPv74Y9WqVUudO3dWjx49VL9+fa1evfq6tt26dWu9//77eu+999S0aVM988wzio2N1QMPPGC2iY2N1fHjxxUeHl6mh1adnZ21du1anT9/Xm3bttWoUaPMu8wqwmHvxo0b67XXXtPChQvVokUL7d69u8TPzyktDw8Pbd++XXXr1tXdd9+txo0ba+TIkbp06dIV/9orr9fuejhyv15Ohw4d9MYbb2jevHlq0aKFPvvsM02aNKnE78MnnnhCzs7OioyMVJ06da56LUxpLF68WD169Cj2oZKDBg3SV199JV9fX33wwQf65JNP1LJlS3Xr1s3u7qTXX39d99xzjx577DE1atRIo0eP1oULFyRJwcHB2rFjh/Lz89WrVy81a9ZMEydOlI+Pj5yciv+VtHTpUo0YMUJTpkxRRESEBgwYoD179qhu3bpXHcvZs2fVunVrDR8+XI8//rj8/f3t2rz88suKi4tTSEiIWrVqJUmKjo7WunXr9Pnnn6tt27b605/+pFdeeUWhoaEl2pflpfAzPygoSO3atdOePXv0wQcfKCoq6rLLPPXUU2rdurWio6MVFRWlwMDAK34TQEREhLZs2aJVq1ZpypQppf4MqchsxtUu0gCqkB07dqhjx446cuSI3QXcQHkbPXq0vvnmG33xxReO7goAcQ0RqriPPvpINWrUUMOGDXXkyBFNmDBBHTp0IAyh3L300kvq2bOnPD09tXHjRr399tt67bXXHN0tAP8fgQhV2rlz5zRt2jSdPHlSfn5+6tGjh92Ts4Hysnv3bs2ZM0fnzp1T/fr19Y9//MO8AwiA43HKDAAAWB4XVQMAAMsjEAEAAMsjEAEAAMsjEAEAAMsjEAGApGXLlsnHx8fR3QDgIAQiAJVWYmKinJ2d1a9fvxItFxYWpvnz59uV3Xvvvfr2229vYO8AVCYEIgCV1uLFizV+/Hht375dP/7443Wtq3r16kW+1gGAdRCIAFRK58+f1+rVqzVmzBj169dPy5Yts6v/9NNP1bZtW7m7u8vPz08DBw6U9NuXv544cUKTJk2SzWYzv2C1uFNmr7/+usLDw+Xq6qqIiAgtX77crt5ms+mtt97SwIED5eHhoYYNG+qTTz4pszEDKDsEIgCV0vvvv69GjRopIiJC999/v5YsWaLC58yuX79eAwcOVN++ffXf//5X8fHxuu222yRJa9as0c0336zY2FilpqYqNTW12PV/9NFHmjBhgqZMmaKDBw/qkUce0YMPPqitW7fatZs1a5YGDx6sAwcOqG/fvho2bJjOnDlTtoMHcMPxpGoAlVKHDh00ePBgTZgwQXl5eQoKCjK/4fv2229X/fr19e677xa7bFhYmCZOnKiJEyeaZcuWLdPEiROVkZFhrr9JkyZatGiR2Wbw4MG6cOGC1q9fL+m3I0RPPfWUnn32WUnShQsXVKNGDW3cuFG9e/cum4EDKBMcIQJQ6aSkpGj37t0aOnSoJKlatWq69957tXjxYklSUlKSunfvfl3bSE5OVocOHezKOnTooOTkZLuy5s2bmz97enrKy8tLp0+fvq5tAyh/fLkrgEpn8eLFysvLU3BwsFlmGIbc3Nz06quvqnr16uXWFxcXF7t5m82mgoKCcts+gBuDI0QAKpW8vDy98847evnll5WUlGRO+/fvV3BwsFatWqXmzZsrPj7+sutwdXVVfn7+FbfTuHFj7dixw65sx44dioyMvCHjAFCxcIQIQKWybt06nT17ViNHjpS3t7dd3aBBg7R48WLNnTtX3bt3V3h4uIYMGaK8vDxt2LBB06ZNk/TbNUTbt2/XkCFD5ObmJj8/vyLbmTp1qgYPHqxWrVqpR48e+vTTT7VmzRpt3ry5XMYJoHxxhAhApbJ48WL16NGjSBiSfgtEX331lXx9ffXBBx/ok08+UcuWLdWtWzft3r3bbBcbG6vjx48rPDxcderUKXY7AwYM0IIFC/TSSy+pSZMmevPNN7V06VJFRUWV1dAAOBB3mQEAAMvjCBEAALA8AhEAALA8AhEAALA8AhEAALA8AhEAALA8AhEAALA8AhEAALA8AhEAALA8AhEAALA8AhEAALA8AhEAALA8AhEAALC8/wcSr7OGMHGgXQAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "action_list = [0,1,2,3,4]\n",
    "action_item = ('Do nothing', 'Turn left', 'Turn right','Accelerate','Brake')\n",
    "h1= []\n",
    "h2 =[]\n",
    "h3 =[]\n",
    "\n",
    "for i in action_list:\n",
    "    h1.append(df[df['PPO_ac']==i]['PPO_ac'].count())\n",
    "    h2.append(df[df['DDPG_ac']==i]['DDPG_ac'].count())\n",
    "    h3.append(df[df['DQN_ac']==i]['DQN_ac'].count())\n",
    "x_labels = np.array(action_list)\n",
    "bar_width = 0.35\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "\n",
    "ax.bar(x_labels - bar_width/2, h1, bar_width, label='PPO agent',color='b')\n",
    "ax.bar(x_labels + bar_width/2, h2, bar_width, label='DDPG agent',color='y')\n",
    "ax.bar(x_labels - bar_width-0.2/2, h3, bar_width, label='DQN agent',color='k')\n",
    "\n",
    "ax.set_xlabel('Action')\n",
    "ax.set_ylabel('Count')\n",
    "ax.set_title('Counts of Actions')\n",
    "plt.legend('lower center')\n",
    "plt.xticks(x_labels,action_item)\n",
    "\n",
    "plt.savefig((p_model/'Test_action_counts2.png'))\n",
    "\n",
    "plt.show()\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Second run -- reshape\n",
    "## Folders and parameters\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Delet previous agent"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2181"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import gc\n",
    "\n",
    "del agent_DQN\n",
    "del agent_DDPG\n",
    "del agent_PPO\n",
    "torch.cuda.empty_cache()\n",
    "gc.collect()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Reload agent"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Max training time steps: 900000\t \n",
      " State dimension: (96, 96)\t \n",
      " Action size: 5\t\n",
      "K-epocs:30\n",
      " max timesteps in one episode: 1500 \n",
      " Expected episode: 1200\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from utility import Config_DQN\n",
    "from utility import Config_DDPG\n",
    "from utility import Config_PPO\n",
    "\n",
    "# Episode = 1000         # test episode\n",
    "# total_test_episodes = Episode\n",
    "Time_stamp = 1500   # time stamp -- simply initial parameters, no use in the test\n",
    "\n",
    "cfg_dqn = Config_DQN(print_ep=1, freq_save_epi=10, n_episodes=10, GAMMA=0.95, SEED=32, eps_end=0.001, max_t=1200)\n",
    "setattr(cfg_dqn, 'state_size', observation_space)\n",
    "setattr(cfg_dqn, 'action_size', action_space) \n",
    "\n",
    "cfg_ppo = Config_PPO(max_training_timesteps=int(900000),max_ep_len=Time_stamp,K_epochs=30, save_model_freq=500,eps_clip=0.2,gamma=0.95)\n",
    "setattr(cfg_ppo, 'state_dim', env.observation_space.shape)\n",
    "setattr(cfg_ppo, 'action_dim', env.action_space.n)\n",
    "\n",
    "print('Max training time steps: {}\\t \\n State dimension: {}\\t \\n Action size: {}\\t'.format(cfg_ppo.max_training_timesteps,cfg_ppo.state_dim,cfg_ppo.action_dim))\n",
    "print(\"K-epocs:{}\\n max timesteps in one episode: {} \\n Expected episode: {}\\n\".format(cfg_ppo.K_epochs,cfg_ppo.max_ep_len,(cfg_ppo.max_training_timesteps // cfg_ppo.max_ep_len)*2))\n",
    "\n",
    "cfg_ddpg = Config_DDPG(Episode=int(Episode), max_ep_len=Time_stamp,random_seed=int(32),freq_save_epi=int(5))\n",
    "setattr(cfg_ddpg, 'state_size', observation_space)  # create the state size parameter\n",
    "setattr(cfg_ddpg, 'action_size', action_space)      # create the action size parameter\n",
    "\n",
    "\n",
    "# Create agent\n",
    "agent_DQN = Agent(cfg_dqn)\n",
    "agent_DDPG = Agent3(cfg_ddpg)\n",
    "agent_PPO = PPO(cfg_ppo)\n",
    "\n",
    "# create folder for the test data\n",
    "file_name = 'Test_Reshape_NearnessII'\n",
    "p_model = (Path()/ 'document'/'Trail_Reshape'/file_name)\n",
    "p_model.mkdir(parents=True, exist_ok=True)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Walk through folders"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Folder list in Experiment:[WindowsPath('Model/DDPG'), WindowsPath('Model/DQN'), WindowsPath('Model/PPO')]\n",
      "\n",
      "DQN folders:[WindowsPath('Model/DQN/DQN_reward_default'), WindowsPath('Model/DQN/DQN_reward_reshape')]\n",
      "\n",
      "DDPG folders:[WindowsPath('Model/DDPG/DDPG_default'), WindowsPath('Model/DDPG/DDPG_reshape_timeout')]\n",
      "\n",
      "PPO folders:[WindowsPath('Model/PPO/PPO_reshape_timeout'), WindowsPath('Model/PPO/PPO_reward_deafult')]\n",
      "\n",
      "------------------------------------\n",
      "\n",
      "DQN model files [WindowsPath('Model/DQN/DQN_reward_default/DQN_local1000.pth'), WindowsPath('Model/DQN/DQN_reward_default/DQN_local500.pth')]\n",
      "DDPG model files [WindowsPath('Model/DDPG/DDPG_default/DDPG_actor1000.pth'), WindowsPath('Model/DDPG/DDPG_default/DDPG_actor500.pth'), WindowsPath('Model/DDPG/DDPG_default/DDPG_critic1000.pth'), WindowsPath('Model/DDPG/DDPG_default/DDPG_critic500.pth')]\n",
      "PPO model files [WindowsPath('Model/PPO/PPO_reshape_timeout/PPO_adreward_1000.pth'), WindowsPath('Model/PPO/PPO_reshape_timeout/PPO_adreward_1018.pth'), WindowsPath('Model/PPO/PPO_reshape_timeout/PPO_adreward_500.pth')]\n"
     ]
    }
   ],
   "source": [
    "from pathlib import Path\n",
    "import pandas as pd\n",
    "import glob\n",
    "import os\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import math\n",
    "# All folders\n",
    "Experiment = (Path() / 'Model')\n",
    "directory = os.walk(Experiment)\n",
    "# filter out the touching the near zoen whenever the car is about to tune angle.\n",
    "gap_threshold = 35\n",
    "folder_list = [x for x in Experiment.iterdir() if x.is_dir()]\n",
    "print(f\"Folder list in Experiment:{folder_list}\\n\")\n",
    "\n",
    "# all sub folders\n",
    "DQN_folders = [a for a in folder_list[1].iterdir() if a.is_dir()]\n",
    "DDPG_folders = [b for b in folder_list[0].iterdir() if b.is_dir()]\n",
    "PPO_folders = [c for c in folder_list[2].iterdir() if c.is_dir()]\n",
    "\n",
    "# default\n",
    "print(f\"DQN folders:{DQN_folders}\\n\")\n",
    "print(f\"DDPG folders:{DDPG_folders}\\n\")\n",
    "print(f\"PPO folders:{PPO_folders}\\n\")\n",
    "print(f\"------------------------------------\\n\")\n",
    "print(f\"DQN model files {list(DQN_folders[0].rglob('*.pth'))}\")\n",
    "print(f\"DDPG model files {list(DDPG_folders[0].rglob('*.pth'))}\")\n",
    "print(f\"PPO model files {list(PPO_folders[0].rglob('*.pth'))}\")\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Load reshape agent"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DQN folders:[WindowsPath('Model/DQN/DQN_reward_default'), WindowsPath('Model/DQN/DQN_reward_reshape')]\n",
      "\n",
      "DDPG folders:[WindowsPath('Model/DDPG/DDPG_default'), WindowsPath('Model/DDPG/DDPG_reshape_timeout')]\n",
      "\n",
      "PPO folders:[WindowsPath('Model/PPO/PPO_reshape_timeout'), WindowsPath('Model/PPO/PPO_reward_deafult')]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "\n",
    "# check folder\n",
    "print(f\"DQN folders:{DQN_folders}\\n\")\n",
    "print(f\"DDPG folders:{DDPG_folders}\\n\")\n",
    "print(f\"PPO folders:{PPO_folders}\\n\")\n",
    "\n",
    "# # Reshape\n",
    "R_DQN_model_path = DQN_folders[0]\n",
    "R_DDPG_model_path = DDPG_folders[0] # :[WindowsPath('Experiment/DDPG/DDPG_default'), WindowsPath('Experiment/DDPG/DDPG_reshape'), WindowsPath('Experiment/DDPG/DDPG_reshape_timeout')]\n",
    "R_PPO_model_path = PPO_folders[0]\n",
    "\n",
    "# device = 'cuda:0' if torch.cuda.is_available() else 'cpu' # device check for GPU avaialbility\n",
    "\n",
    "# Reshape Agent\n",
    "DQN_model_name = f\"DQN_local{1000}.pth\"\n",
    "agent_DQN.qnetwork_local.load_state_dict(torch.load((R_DQN_model_path/DQN_model_name),map_location=torch.device('cuda')))\n",
    "\n",
    "actor_name = f'DDPG_actor{1000}.pth'\n",
    "critic_name = f'DDPG_critic{1000}.pth'\n",
    "agent_DDPG.actor_local.load_state_dict(torch.load((R_DDPG_model_path/actor_name),map_location=torch.device('cuda')))\n",
    "agent_DDPG.critic_local.load_state_dict(torch.load((R_DDPG_model_path/critic_name),map_location=torch.device('cuda')))\n",
    "PPOmodel_name = f\"PPO_adreward_{1018}.pth\"\n",
    "agent_PPO.load((R_PPO_model_path / PPOmodel_name))\n",
    "\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Test start"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## PPO Reshape agent"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Episode: 1 \t\t Reward: 154.0000000000029\n",
      "Episode: 2 \t\t Reward: 153.5000000000029\n",
      "Episode: 3 \t\t Reward: 154.66666666666956\n",
      "Total test time: 00:00:32.53\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from collections import deque\n",
    "import numpy as np\n",
    "from utility import get_unique_numbers\n",
    "import time\n",
    "\n",
    "test_running_reward = 0\n",
    "# total_test_episodes = 3\n",
    "ac =[]\n",
    "start = time.time()\n",
    "runn_ep =[]\n",
    "scores =[]\n",
    "scores_deque = deque(maxlen=100)\n",
    "avg_scores =[]\n",
    "Bmp =[]\n",
    "T_Bmp = []\n",
    "if p_model.is_dir():\n",
    "    nb_name = 'PPO_Test_log.txt'\n",
    "with open(p_model.joinpath(nb_name), \"w\") as f:\n",
    "\n",
    "    for ep in range(1, total_test_episodes+1):\n",
    "        start_ep = time.time()\n",
    "        ep_reward = 0\n",
    "        runn_ep.append(ep)\n",
    "        state = env.reset()\n",
    "        Ep_bump =[]\n",
    "        for t in range(1, 1500):\n",
    "            action = agent_PPO.select_action(state)\n",
    "            ac.append(action)\n",
    "            state, reward, done, bumps = env.step(action)\n",
    "            Ep_bump.append(bumps)\n",
    "            T_Bmp.append(bumps)\n",
    "            ep_reward += reward\n",
    "            # env.render('rgb_array') # uncommand to display -- warning: expensive resources.\n",
    "            if done:\n",
    "                break\n",
    "        Bmp.append(Ep_bump)\n",
    "        scores_deque.append(ep_reward)\n",
    "        avg_scores.append(np.mean(scores_deque))\n",
    "        end_ep = time.time()\n",
    "        hours_ep, rem_ep = divmod(end_ep-start_ep, 3600)\n",
    "        minutes_ep, seconds_ep = divmod(rem_ep, 60)\n",
    "\n",
    "        f.write(\"Episode{} (survive time): {:0>2}:{:0>2}:{:05.2f}\\n\".format(ep,int(hours_ep),int(minutes_ep),seconds_ep))\n",
    "        test_running_reward +=  ep_reward\n",
    "\n",
    "        # print('Episode: {} \\t\\t Reward: {}'.format(ep, round(ep_reward, 2)))\n",
    "        print('Episode: {} \\t\\t Reward: {}'.format(ep, avg_scores[ep-1]))\n",
    "\n",
    "        scores.append(round(ep_reward, 2))\n",
    "        ep_reward = 0\n",
    "    env.close()\n",
    "\n",
    "    end = time.time()\n",
    "    hours, rem = divmod(end-start, 3600)\n",
    "    minutes, seconds = divmod(rem, 60)\n",
    "    print(\"Total test time: {:0>2}:{:0>2}:{:05.2f}\".format(int(hours),int(minutes),seconds))\n",
    "    \n",
    "    # count the bump instances.\n",
    "    count_list_PPO = []\n",
    "    for i in range(len(Bmp)):\n",
    "        count_list_PPO.append(count_impulses(Bmp[i]))\n",
    "    data_PPO = {'Bump_instance': count_list_PPO}\n",
    "    Bump_track = pd.DataFrame(data_PPO)\n",
    "    Tcount_PPO = count_edgezone(T_Bmp,gap_threshold)\n",
    "\n",
    "\n",
    "    f.write(\"Total Test time (survive time): {:0>2}:{:0>2}:{:05.2f}\\n\".format(int(hours),int(minutes),seconds))\n",
    "    f.write('Toal epoch:%s\\n'%total_test_episodes)\n",
    "    f.write(\"Bump instances of total: %s\\n\"%Tcount_PPO)\n",
    "    f.write(\"Bump to edge zone %s\\n\"%count_edgezone(T_Bmp, gap_threshold))\n",
    "    f.close()\n",
    "\n",
    "PPO_avg = avg_scores\n",
    "PPO_scores = scores\n",
    "PPO_ac = ac\n",
    "\n",
    "csv_name = f\"PPO_reshape\"\n",
    "# score_save(scores,avg_scores,file_name=csv_name,path=p_model,print_=False) # save score table as csv file\n",
    "score_action_save(ac,scores,avg_scores,file_name=csv_name,path=p_model,print_=False) # save score table as csv file\n",
    "count_impulse_file = f\"PPO_reshape_count\"\n",
    "Bump_track.to_csv((p_model/f\"{count_impulse_file}.csv\")) # save bump instances as csv file\n",
    "\n",
    "# count the bump instances.\n",
    "count_list_PPO = []\n",
    "for i in range(len(Bmp)):\n",
    "    count_list_PPO.append(count_impulses(Bmp[i]))\n",
    "data_PPO = {'Bump_instance': count_list_PPO}\n",
    "Bump_track_PPO = pd.DataFrame(data_PPO)\n",
    "\n",
    "# fig = plt.figure()\n",
    "# plt.plot(T_Bmp)\n",
    "# plt.title(f'Total bump instances:{count_impulses(T_Bmp)}')\n",
    "\n",
    "# # plot the bump instances.\n",
    "# for i in range(5):\n",
    "#     plt.subplot(2, 5, i+1)\n",
    "#     plt.plot(Bmp[i])\n",
    "#     plt.title(f'Ct_:{count_list[i]}')\n",
    "#     plt.subplot(2, 5, i+6)\n",
    "#     plt.plot(Bmp[i+5])\n",
    "#     plt.title(f'Ct_:{count_list[i+5]}')\n",
    "\n",
    "# fig = plt.figure()\n",
    "# plt.plot(T_Bmp)\n",
    "# plt.title(f'Total bump instances:{Tcount_PPO}')\n",
    "\n",
    "\n",
    "fig = plt.figure()\n",
    "ax = fig.add_subplot(111)\n",
    "plt.plot(avg_scores)\n",
    "# plt.plot(df['Averagescores'])\n",
    "plt.ylabel('Scores')\n",
    "plt.xlabel(f'Episode {len(avg_scores)}')\n",
    "plt.show()\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "\n",
    "action_list = [0,1,2,3,4]\n",
    "h=[]\n",
    "for i in action_list:\n",
    "    h.append(ac.count(i))\n",
    "bar_contain = plt.bar(action_list,h)\n",
    "plt.xlabel('Action items')\n",
    "plt.ylabel('Number of action item used during trial')\n",
    "plt.bar_label(bar_contain)\n",
    "plt.show()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## DDPG Reshape agent"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import gc\n",
    "\n",
    "# del agent_DQN\n",
    "# del agent_DDPG\n",
    "del agent_PPO\n",
    "torch.cuda.empty_cache()\n",
    "gc.collect()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Episode: 1 \t\t Reward: -72.40000000000006\n",
      "Episode: 2 \t\t Reward: -275.4999999999993\n",
      "Episode: 3 \t\t Reward: -372.0333333333329\n",
      "Episode: 4 \t\t Reward: -424.6124999999997\n",
      "Episode: 5 \t\t Reward: -333.8499999999994\n",
      "Episode: 6 \t\t Reward: -270.98333333333244\n",
      "Episode: 7 \t\t Reward: -241.82142857142782\n",
      "Episode: 8 \t\t Reward: -209.65624999999918\n",
      "Episode: 9 \t\t Reward: -246.3388888888882\n",
      "Episode: 10 \t\t Reward: -225.03999999999942\n",
      "Episode: 11 \t\t Reward: -210.34545454545403\n",
      "Episode: 12 \t\t Reward: -237.46666666666624\n",
      "Episode: 13 \t\t Reward: -222.72307692307655\n",
      "Episode: 14 \t\t Reward: -211.23214285714252\n",
      "Episode: 15 \t\t Reward: -234.94666666666637\n",
      "Episode: 16 \t\t Reward: -220.86874999999966\n",
      "Episode: 17 \t\t Reward: -240.8735294117644\n",
      "Episode: 18 \t\t Reward: -229.55277777777752\n",
      "Episode: 19 \t\t Reward: -220.62105263157872\n",
      "Episode: 20 \t\t Reward: -210.15749999999974\n",
      "Episode: 21 \t\t Reward: -201.3285714285712\n",
      "Episode: 22 \t\t Reward: -216.6840909090907\n",
      "Episode: 23 \t\t Reward: -210.88478260869547\n",
      "Episode: 24 \t\t Reward: -202.91458333333313\n",
      "Episode: 25 \t\t Reward: -196.3399999999998\n",
      "Episode: 26 \t\t Reward: -189.66923076923058\n",
      "Episode: 27 \t\t Reward: -202.82777777777758\n",
      "Episode: 28 \t\t Reward: -197.76785714285697\n",
      "Episode: 29 \t\t Reward: -190.44482758620669\n",
      "Episode: 30 \t\t Reward: -185.20999999999984\n",
      "Episode: 31 \t\t Reward: -180.1951612903224\n",
      "Episode: 32 \t\t Reward: -175.02812499999982\n",
      "Episode: 33 \t\t Reward: -168.35606060606037\n",
      "Episode: 34 \t\t Reward: -163.39999999999978\n",
      "Episode: 35 \t\t Reward: -159.44428571428548\n",
      "Episode: 36 \t\t Reward: -154.5652777777775\n",
      "Episode: 37 \t\t Reward: -151.11891891891867\n",
      "Episode: 38 \t\t Reward: -161.99210526315764\n",
      "Episode: 39 \t\t Reward: -153.50512820512787\n",
      "Episode: 40 \t\t Reward: -150.93499999999966\n",
      "Episode: 41 \t\t Reward: -147.93170731707286\n",
      "Episode: 42 \t\t Reward: -143.02976190476153\n",
      "Episode: 43 \t\t Reward: -140.84883720930196\n",
      "Episode: 44 \t\t Reward: -139.25681818181783\n",
      "Episode: 45 \t\t Reward: -137.12888888888855\n",
      "Episode: 46 \t\t Reward: -134.5510869565214\n",
      "Episode: 47 \t\t Reward: -132.13191489361668\n",
      "Episode: 48 \t\t Reward: -130.07916666666634\n",
      "Episode: 49 \t\t Reward: -127.42142857142827\n",
      "Episode: 50 \t\t Reward: -124.9609999999997\n",
      "Episode: 51 \t\t Reward: -133.3872549019605\n",
      "Episode: 52 \t\t Reward: -132.11923076923048\n",
      "Episode: 53 \t\t Reward: -130.58867924528275\n",
      "Episode: 54 \t\t Reward: -138.0916666666664\n",
      "Episode: 55 \t\t Reward: -135.57818181818158\n",
      "Episode: 56 \t\t Reward: -132.90089285714257\n",
      "Episode: 57 \t\t Reward: -131.514035087719\n",
      "Episode: 58 \t\t Reward: -130.20258620689626\n",
      "Episode: 59 \t\t Reward: -129.02457627118616\n",
      "Episode: 60 \t\t Reward: -126.8716666666664\n",
      "Episode: 61 \t\t Reward: -125.76147540983581\n",
      "Episode: 62 \t\t Reward: -124.59838709677395\n",
      "Episode: 63 \t\t Reward: -123.62857142857119\n",
      "Episode: 64 \t\t Reward: -122.67031249999977\n",
      "Episode: 65 \t\t Reward: -122.05923076923054\n",
      "Episode: 66 \t\t Reward: -121.34393939393917\n",
      "Episode: 67 \t\t Reward: -120.26716417910426\n",
      "Episode: 68 \t\t Reward: -127.1551470588233\n",
      "Episode: 69 \t\t Reward: -126.29347826086935\n",
      "Episode: 70 \t\t Reward: -124.98214285714265\n",
      "Episode: 71 \t\t Reward: -124.01338028168995\n",
      "Episode: 72 \t\t Reward: -122.77847222222204\n",
      "Episode: 73 \t\t Reward: -121.87945205479433\n",
      "Episode: 74 \t\t Reward: -120.28783783783763\n",
      "Episode: 75 \t\t Reward: -119.16333333333314\n",
      "Episode: 76 \t\t Reward: -118.19210526315771\n",
      "Episode: 77 \t\t Reward: -117.20259740259722\n",
      "Episode: 78 \t\t Reward: -116.09871794871778\n",
      "Episode: 79 \t\t Reward: -122.10126582278463\n",
      "Episode: 80 \t\t Reward: -121.01812499999983\n",
      "Episode: 81 \t\t Reward: -126.70061728395044\n",
      "Episode: 82 \t\t Reward: -132.3878048780486\n",
      "Episode: 83 \t\t Reward: -131.13795180722872\n",
      "Episode: 84 \t\t Reward: -130.11666666666648\n",
      "Episode: 85 \t\t Reward: -128.85764705882335\n",
      "Episode: 86 \t\t Reward: -127.82034883720912\n",
      "Episode: 87 \t\t Reward: -133.1448275862067\n",
      "Episode: 88 \t\t Reward: -132.56477272727258\n",
      "Episode: 89 \t\t Reward: -131.8842696629212\n",
      "Episode: 90 \t\t Reward: -131.12277777777766\n",
      "Episode: 91 \t\t Reward: -129.15274725274708\n",
      "Episode: 92 \t\t Reward: -133.9239130434781\n",
      "Episode: 93 \t\t Reward: -132.8413978494622\n",
      "Episode: 94 \t\t Reward: -132.02234042553175\n",
      "Episode: 95 \t\t Reward: -130.95210526315776\n",
      "Episode: 96 \t\t Reward: -130.12968749999987\n",
      "Episode: 97 \t\t Reward: -129.6463917525772\n",
      "Episode: 98 \t\t Reward: -128.86785714285702\n",
      "Episode: 99 \t\t Reward: -128.21868686868675\n",
      "Episode: 100 \t\t Reward: -127.5824999999999\n",
      "Episode: 101 \t\t Reward: -132.66049999999987\n",
      "Episode: 102 \t\t Reward: -133.36899999999991\n",
      "Episode: 103 \t\t Reward: -128.3194999999999\n",
      "Episode: 104 \t\t Reward: -122.46899999999988\n",
      "Episode: 105 \t\t Reward: -122.56649999999989\n",
      "Episode: 106 \t\t Reward: -123.51499999999992\n",
      "Episode: 107 \t\t Reward: -128.72099999999992\n",
      "Episode: 108 \t\t Reward: -134.75999999999993\n",
      "Episode: 109 \t\t Reward: -129.9619999999999\n",
      "Episode: 110 \t\t Reward: -129.4824999999999\n",
      "Episode: 111 \t\t Reward: -128.66749999999988\n",
      "Episode: 112 \t\t Reward: -123.30799999999986\n",
      "Episode: 113 \t\t Reward: -123.13899999999987\n",
      "Episode: 114 \t\t Reward: -123.16349999999986\n",
      "Episode: 115 \t\t Reward: -117.63999999999987\n",
      "Episode: 116 \t\t Reward: -118.14649999999988\n",
      "Episode: 117 \t\t Reward: -113.1919999999999\n",
      "Episode: 118 \t\t Reward: -112.97249999999987\n",
      "Episode: 119 \t\t Reward: -113.21399999999987\n",
      "Episode: 120 \t\t Reward: -113.09899999999989\n",
      "Episode: 121 \t\t Reward: -113.00349999999987\n",
      "Episode: 122 \t\t Reward: -107.61049999999986\n",
      "Episode: 123 \t\t Reward: -107.54099999999988\n",
      "Episode: 124 \t\t Reward: -107.97499999999987\n",
      "Episode: 125 \t\t Reward: -107.99649999999988\n",
      "Episode: 126 \t\t Reward: -108.2634999999999\n",
      "Episode: 127 \t\t Reward: -103.6634999999999\n",
      "Episode: 128 \t\t Reward: -108.92849999999987\n",
      "Episode: 129 \t\t Reward: -109.5264999999999\n",
      "Episode: 130 \t\t Reward: -108.9509999999999\n",
      "Episode: 131 \t\t Reward: -114.49649999999987\n",
      "Episode: 132 \t\t Reward: -114.63299999999988\n",
      "Episode: 133 \t\t Reward: -115.5819999999999\n",
      "Episode: 134 \t\t Reward: -115.7659999999999\n",
      "Episode: 135 \t\t Reward: -116.1939999999999\n",
      "Episode: 136 \t\t Reward: -117.06999999999991\n",
      "Episode: 137 \t\t Reward: -117.38199999999992\n",
      "Episode: 138 \t\t Reward: -116.29899999999991\n",
      "Episode: 139 \t\t Reward: -118.89149999999994\n",
      "Episode: 140 \t\t Reward: -116.58449999999989\n",
      "Episode: 141 \t\t Reward: -116.99299999999988\n",
      "Episode: 142 \t\t Reward: -117.6939999999999\n",
      "Episode: 143 \t\t Reward: -117.1344999999999\n",
      "Episode: 144 \t\t Reward: -116.9729999999999\n",
      "Episode: 145 \t\t Reward: -116.53899999999989\n",
      "Episode: 146 \t\t Reward: -117.0584999999999\n",
      "Episode: 147 \t\t Reward: -117.37499999999991\n",
      "Episode: 148 \t\t Reward: -122.8429999999999\n",
      "Episode: 149 \t\t Reward: -128.14749999999992\n",
      "Episode: 150 \t\t Reward: -133.82599999999994\n",
      "Episode: 151 \t\t Reward: -128.71599999999992\n",
      "Episode: 152 \t\t Reward: -128.52099999999993\n",
      "Episode: 153 \t\t Reward: -128.90199999999993\n",
      "Episode: 154 \t\t Reward: -124.10399999999994\n",
      "Episode: 155 \t\t Reward: -124.77149999999995\n",
      "Episode: 156 \t\t Reward: -124.86049999999994\n",
      "Episode: 157 \t\t Reward: -124.81249999999994\n",
      "Episode: 158 \t\t Reward: -125.08299999999991\n",
      "Episode: 159 \t\t Reward: -124.80699999999993\n",
      "Episode: 160 \t\t Reward: -124.45499999999991\n",
      "Episode: 161 \t\t Reward: -123.81049999999988\n",
      "Episode: 162 \t\t Reward: -129.10999999999987\n",
      "Episode: 163 \t\t Reward: -128.75549999999987\n",
      "Episode: 164 \t\t Reward: -128.1309999999999\n",
      "Episode: 165 \t\t Reward: -127.29999999999987\n",
      "Episode: 166 \t\t Reward: -132.20249999999987\n",
      "Episode: 167 \t\t Reward: -131.00199999999984\n",
      "Episode: 168 \t\t Reward: -130.70699999999982\n",
      "Episode: 169 \t\t Reward: -130.49099999999984\n",
      "Episode: 170 \t\t Reward: -130.41199999999984\n",
      "Episode: 171 \t\t Reward: -135.57049999999984\n",
      "Episode: 172 \t\t Reward: -135.60449999999983\n",
      "Episode: 173 \t\t Reward: -140.7434999999998\n",
      "Episode: 174 \t\t Reward: -141.59349999999984\n",
      "Episode: 175 \t\t Reward: -141.62849999999983\n",
      "Episode: 176 \t\t Reward: -145.8444999999998\n",
      "Episode: 177 \t\t Reward: -143.71449999999982\n",
      "Episode: 178 \t\t Reward: -149.34399999999977\n",
      "Episode: 179 \t\t Reward: -144.09449999999978\n",
      "Episode: 180 \t\t Reward: -143.29149999999976\n",
      "Episode: 181 \t\t Reward: -137.61499999999975\n",
      "Episode: 182 \t\t Reward: -131.98049999999975\n",
      "Episode: 183 \t\t Reward: -131.84649999999976\n",
      "Episode: 184 \t\t Reward: -131.10199999999975\n",
      "Episode: 185 \t\t Reward: -130.61399999999972\n",
      "Episode: 186 \t\t Reward: -130.47999999999973\n",
      "Episode: 187 \t\t Reward: -122.93949999999967\n",
      "Episode: 188 \t\t Reward: -127.04699999999966\n",
      "Episode: 189 \t\t Reward: -126.86699999999968\n",
      "Episode: 190 \t\t Reward: -132.15799999999967\n",
      "Episode: 191 \t\t Reward: -132.7019999999997\n",
      "Episode: 192 \t\t Reward: -127.38899999999973\n",
      "Episode: 193 \t\t Reward: -127.58049999999972\n",
      "Episode: 194 \t\t Reward: -127.81399999999971\n",
      "Episode: 195 \t\t Reward: -133.3129999999997\n",
      "Episode: 196 \t\t Reward: -133.2809999999997\n",
      "Episode: 197 \t\t Reward: -137.79399999999973\n",
      "Episode: 198 \t\t Reward: -137.9379999999997\n",
      "Episode: 199 \t\t Reward: -135.59199999999967\n",
      "Episode: 200 \t\t Reward: -135.49349999999967\n",
      "Episode: 201 \t\t Reward: -135.09199999999967\n",
      "Episode: 202 \t\t Reward: -129.89899999999966\n",
      "Episode: 203 \t\t Reward: -129.92399999999967\n",
      "Episode: 204 \t\t Reward: -130.30799999999968\n",
      "Episode: 205 \t\t Reward: -131.28399999999968\n",
      "Episode: 206 \t\t Reward: -131.2674999999997\n",
      "Episode: 207 \t\t Reward: -125.79599999999971\n",
      "Episode: 208 \t\t Reward: -119.91749999999969\n",
      "Episode: 209 \t\t Reward: -124.5829999999997\n",
      "Episode: 210 \t\t Reward: -124.72749999999971\n",
      "Episode: 211 \t\t Reward: -125.42949999999972\n",
      "Episode: 212 \t\t Reward: -126.14749999999972\n",
      "Episode: 213 \t\t Reward: -126.25899999999973\n",
      "Episode: 214 \t\t Reward: -126.11549999999974\n",
      "Episode: 215 \t\t Reward: -131.57199999999975\n",
      "Episode: 216 \t\t Reward: -130.61749999999972\n",
      "Episode: 217 \t\t Reward: -130.6384999999997\n",
      "Episode: 218 \t\t Reward: -130.4029999999997\n",
      "Episode: 219 \t\t Reward: -129.6929999999997\n",
      "Episode: 220 \t\t Reward: -129.9984999999997\n",
      "Episode: 221 \t\t Reward: -130.0614999999997\n",
      "Episode: 222 \t\t Reward: -135.35749999999973\n",
      "Episode: 223 \t\t Reward: -132.89399999999966\n",
      "Episode: 224 \t\t Reward: -132.89949999999968\n",
      "Episode: 225 \t\t Reward: -132.85349999999966\n",
      "Episode: 226 \t\t Reward: -133.12799999999964\n",
      "Episode: 227 \t\t Reward: -132.70149999999964\n",
      "Episode: 228 \t\t Reward: -127.56849999999967\n",
      "Episode: 229 \t\t Reward: -127.59599999999968\n",
      "Episode: 230 \t\t Reward: -127.7129999999997\n",
      "Episode: 231 \t\t Reward: -122.48549999999969\n",
      "Episode: 232 \t\t Reward: -121.89049999999966\n",
      "Episode: 233 \t\t Reward: -121.47299999999964\n",
      "Episode: 234 \t\t Reward: -120.94549999999963\n",
      "Episode: 235 \t\t Reward: -121.01999999999964\n",
      "Episode: 236 \t\t Reward: -120.86999999999964\n",
      "Episode: 237 \t\t Reward: -126.00349999999963\n",
      "Episode: 238 \t\t Reward: -126.59099999999965\n",
      "Episode: 239 \t\t Reward: -126.02699999999965\n",
      "Episode: 240 \t\t Reward: -128.2739999999997\n",
      "Episode: 241 \t\t Reward: -128.4284999999997\n",
      "Episode: 242 \t\t Reward: -128.8279999999997\n",
      "Episode: 243 \t\t Reward: -129.3369999999997\n",
      "Episode: 244 \t\t Reward: -134.4439999999997\n",
      "Episode: 245 \t\t Reward: -135.1829999999997\n",
      "Episode: 246 \t\t Reward: -139.62699999999973\n",
      "Episode: 247 \t\t Reward: -137.46199999999968\n",
      "Episode: 248 \t\t Reward: -132.18549999999965\n",
      "Episode: 249 \t\t Reward: -127.71949999999966\n",
      "Episode: 250 \t\t Reward: -122.86199999999968\n",
      "Episode: 251 \t\t Reward: -122.35449999999966\n",
      "Episode: 252 \t\t Reward: -127.04349999999965\n",
      "Episode: 253 \t\t Reward: -126.37849999999968\n",
      "Episode: 254 \t\t Reward: -131.49849999999967\n",
      "Episode: 255 \t\t Reward: -130.45299999999963\n",
      "Episode: 256 \t\t Reward: -130.1054999999996\n",
      "Episode: 257 \t\t Reward: -130.29149999999962\n",
      "Episode: 258 \t\t Reward: -130.1794999999996\n",
      "Episode: 259 \t\t Reward: -129.8859999999996\n",
      "Episode: 260 \t\t Reward: -130.6659999999996\n",
      "Episode: 261 \t\t Reward: -130.71749999999963\n",
      "Episode: 262 \t\t Reward: -130.32849999999965\n",
      "Episode: 263 \t\t Reward: -130.53549999999964\n",
      "Episode: 264 \t\t Reward: -128.72699999999963\n",
      "Episode: 265 \t\t Reward: -129.19049999999962\n",
      "Episode: 266 \t\t Reward: -123.74549999999961\n",
      "Episode: 267 \t\t Reward: -124.75999999999965\n",
      "Episode: 268 \t\t Reward: -119.89849999999966\n",
      "Episode: 269 \t\t Reward: -119.51449999999963\n",
      "Episode: 270 \t\t Reward: -119.50749999999962\n",
      "Episode: 271 \t\t Reward: -113.30599999999963\n",
      "Episode: 272 \t\t Reward: -113.41399999999962\n",
      "Episode: 273 \t\t Reward: -108.33299999999963\n",
      "Episode: 274 \t\t Reward: -107.90049999999962\n",
      "Episode: 275 \t\t Reward: -107.98999999999961\n",
      "Episode: 276 \t\t Reward: -103.31899999999963\n",
      "Episode: 277 \t\t Reward: -105.15299999999966\n",
      "Episode: 278 \t\t Reward: -99.26399999999965\n",
      "Episode: 279 \t\t Reward: -98.34399999999965\n",
      "Episode: 280 \t\t Reward: -98.81149999999967\n",
      "Episode: 281 \t\t Reward: -98.84649999999965\n",
      "Episode: 282 \t\t Reward: -97.95049999999965\n",
      "Episode: 283 \t\t Reward: -98.23549999999963\n",
      "Episode: 284 \t\t Reward: -96.85649999999961\n",
      "Episode: 285 \t\t Reward: -97.92699999999964\n",
      "Episode: 286 \t\t Reward: -103.56699999999964\n",
      "Episode: 287 \t\t Reward: -105.86749999999965\n",
      "Episode: 288 \t\t Reward: -101.70199999999966\n",
      "Episode: 289 \t\t Reward: -100.98349999999964\n",
      "Episode: 290 \t\t Reward: -95.70249999999966\n",
      "Episode: 291 \t\t Reward: -96.02049999999967\n",
      "Episode: 292 \t\t Reward: -96.23449999999968\n",
      "Episode: 293 \t\t Reward: -95.32649999999965\n",
      "Episode: 294 \t\t Reward: -92.87449999999963\n",
      "Episode: 295 \t\t Reward: -87.58249999999961\n",
      "Episode: 296 \t\t Reward: -87.10149999999962\n",
      "Episode: 297 \t\t Reward: -81.9324999999996\n",
      "Episode: 298 \t\t Reward: -81.25349999999959\n",
      "Episode: 299 \t\t Reward: -82.88749999999963\n",
      "Episode: 300 \t\t Reward: -82.33849999999963\n",
      "Episode: 301 \t\t Reward: -76.4679999999996\n",
      "Episode: 302 \t\t Reward: -76.72849999999958\n",
      "Episode: 303 \t\t Reward: -74.46199999999956\n",
      "Episode: 304 \t\t Reward: -74.13349999999954\n",
      "Episode: 305 \t\t Reward: -73.98499999999956\n",
      "Episode: 306 \t\t Reward: -74.04549999999955\n",
      "Episode: 307 \t\t Reward: -79.57899999999955\n",
      "Episode: 308 \t\t Reward: -80.00249999999956\n",
      "Episode: 309 \t\t Reward: -75.35199999999956\n",
      "Episode: 310 \t\t Reward: -75.63099999999956\n",
      "Episode: 311 \t\t Reward: -75.74649999999956\n",
      "Episode: 312 \t\t Reward: -75.37599999999956\n",
      "Episode: 313 \t\t Reward: -75.48049999999955\n",
      "Episode: 314 \t\t Reward: -80.90349999999955\n",
      "Episode: 315 \t\t Reward: -75.97299999999954\n",
      "Episode: 316 \t\t Reward: -76.43599999999957\n",
      "Episode: 317 \t\t Reward: -76.51899999999956\n",
      "Episode: 318 \t\t Reward: -77.03199999999958\n",
      "Episode: 319 \t\t Reward: -77.08349999999959\n",
      "Episode: 320 \t\t Reward: -77.32549999999958\n",
      "Episode: 321 \t\t Reward: -82.8504999999996\n",
      "Episode: 322 \t\t Reward: -77.16799999999955\n",
      "Episode: 323 \t\t Reward: -79.5524999999996\n",
      "Episode: 324 \t\t Reward: -78.97499999999958\n",
      "Episode: 325 \t\t Reward: -79.43149999999957\n",
      "Episode: 326 \t\t Reward: -79.41999999999958\n",
      "Episode: 327 \t\t Reward: -79.43049999999958\n",
      "Episode: 328 \t\t Reward: -79.39899999999957\n",
      "Episode: 329 \t\t Reward: -84.71349999999957\n",
      "Episode: 330 \t\t Reward: -85.4474999999996\n",
      "Episode: 331 \t\t Reward: -85.57949999999961\n",
      "Episode: 332 \t\t Reward: -86.47849999999963\n",
      "Episode: 333 \t\t Reward: -86.93849999999964\n",
      "Episode: 334 \t\t Reward: -88.19499999999964\n",
      "Episode: 335 \t\t Reward: -92.56649999999964\n",
      "Episode: 336 \t\t Reward: -92.77749999999963\n",
      "Episode: 337 \t\t Reward: -87.98649999999965\n",
      "Episode: 338 \t\t Reward: -83.15699999999966\n",
      "Episode: 339 \t\t Reward: -82.32899999999961\n",
      "Episode: 340 \t\t Reward: -82.51599999999962\n",
      "Episode: 341 \t\t Reward: -82.58299999999961\n",
      "Episode: 342 \t\t Reward: -82.61549999999963\n",
      "Episode: 343 \t\t Reward: -82.64249999999961\n",
      "Episode: 344 \t\t Reward: -77.12899999999964\n",
      "Episode: 345 \t\t Reward: -76.38749999999963\n",
      "Episode: 346 \t\t Reward: -71.63449999999963\n",
      "Episode: 347 \t\t Reward: -73.71999999999966\n",
      "Episode: 348 \t\t Reward: -74.02849999999965\n",
      "Episode: 349 \t\t Reward: -73.72099999999966\n",
      "Episode: 350 \t\t Reward: -72.36049999999963\n",
      "Episode: 351 \t\t Reward: -72.10799999999963\n",
      "Episode: 352 \t\t Reward: -67.44249999999963\n",
      "Episode: 353 \t\t Reward: -67.58599999999964\n",
      "Episode: 354 \t\t Reward: -66.3369999999996\n",
      "Episode: 355 \t\t Reward: -67.34949999999964\n",
      "Episode: 356 \t\t Reward: -68.54899999999967\n",
      "Episode: 357 \t\t Reward: -68.28649999999966\n",
      "Episode: 358 \t\t Reward: -67.87599999999966\n",
      "Episode: 359 \t\t Reward: -68.74549999999967\n",
      "Episode: 360 \t\t Reward: -68.85999999999966\n",
      "Episode: 361 \t\t Reward: -69.56999999999967\n",
      "Episode: 362 \t\t Reward: -64.97499999999965\n",
      "Episode: 363 \t\t Reward: -65.30099999999966\n",
      "Episode: 364 \t\t Reward: -67.03149999999968\n",
      "Episode: 365 \t\t Reward: -71.94949999999969\n",
      "Episode: 366 \t\t Reward: -77.04349999999968\n",
      "Episode: 367 \t\t Reward: -77.24799999999969\n",
      "Episode: 368 \t\t Reward: -74.80799999999965\n",
      "Episode: 369 \t\t Reward: -75.39299999999966\n",
      "Episode: 370 \t\t Reward: -74.91849999999965\n",
      "Episode: 371 \t\t Reward: -81.20599999999968\n",
      "Episode: 372 \t\t Reward: -81.28749999999967\n",
      "Episode: 373 \t\t Reward: -80.51849999999963\n",
      "Episode: 374 \t\t Reward: -80.53049999999963\n",
      "Episode: 375 \t\t Reward: -80.32449999999963\n",
      "Episode: 376 \t\t Reward: -85.70099999999964\n",
      "Episode: 377 \t\t Reward: -84.8894999999996\n",
      "Episode: 378 \t\t Reward: -84.94199999999962\n",
      "Episode: 379 \t\t Reward: -85.58499999999964\n",
      "Episode: 380 \t\t Reward: -90.80299999999964\n",
      "Episode: 381 \t\t Reward: -96.25099999999966\n",
      "Episode: 382 \t\t Reward: -96.79099999999966\n",
      "Episode: 383 \t\t Reward: -95.73649999999965\n",
      "Episode: 384 \t\t Reward: -98.00749999999967\n",
      "Episode: 385 \t\t Reward: -97.65999999999967\n",
      "Episode: 386 \t\t Reward: -97.0294999999997\n",
      "Episode: 387 \t\t Reward: -96.9749999999997\n",
      "Episode: 388 \t\t Reward: -96.57149999999969\n",
      "Episode: 389 \t\t Reward: -96.9569999999997\n",
      "Episode: 390 \t\t Reward: -97.1769999999997\n",
      "Episode: 391 \t\t Reward: -96.79499999999969\n",
      "Episode: 392 \t\t Reward: -96.98699999999968\n",
      "Episode: 393 \t\t Reward: -97.69049999999972\n",
      "Episode: 394 \t\t Reward: -105.24749999999975\n",
      "Episode: 395 \t\t Reward: -105.28999999999975\n",
      "Episode: 396 \t\t Reward: -104.83649999999975\n",
      "Episode: 397 \t\t Reward: -104.92799999999976\n",
      "Episode: 398 \t\t Reward: -105.45249999999974\n",
      "Episode: 399 \t\t Reward: -105.51699999999975\n",
      "Episode: 400 \t\t Reward: -106.31149999999973\n",
      "Episode: 401 \t\t Reward: -107.58599999999977\n",
      "Episode: 402 \t\t Reward: -107.68599999999978\n",
      "Episode: 403 \t\t Reward: -109.09849999999979\n",
      "Episode: 404 \t\t Reward: -114.8029999999998\n",
      "Episode: 405 \t\t Reward: -114.95799999999977\n",
      "Episode: 406 \t\t Reward: -120.26299999999978\n",
      "Episode: 407 \t\t Reward: -114.32499999999978\n",
      "Episode: 408 \t\t Reward: -114.43249999999976\n",
      "Episode: 409 \t\t Reward: -114.15799999999977\n",
      "Episode: 410 \t\t Reward: -113.80999999999977\n",
      "Episode: 411 \t\t Reward: -113.17199999999978\n",
      "Episode: 412 \t\t Reward: -113.20749999999978\n",
      "Episode: 413 \t\t Reward: -118.45249999999976\n",
      "Episode: 414 \t\t Reward: -112.81299999999978\n",
      "Episode: 415 \t\t Reward: -112.91299999999976\n",
      "Episode: 416 \t\t Reward: -112.63899999999975\n",
      "Episode: 417 \t\t Reward: -112.22599999999979\n",
      "Episode: 418 \t\t Reward: -117.38949999999977\n",
      "Episode: 419 \t\t Reward: -117.57549999999975\n",
      "Episode: 420 \t\t Reward: -115.28949999999973\n",
      "Episode: 421 \t\t Reward: -109.34299999999972\n",
      "Episode: 422 \t\t Reward: -110.33749999999972\n",
      "Episode: 423 \t\t Reward: -109.67949999999973\n",
      "Episode: 424 \t\t Reward: -109.96649999999975\n",
      "Episode: 425 \t\t Reward: -114.20349999999974\n",
      "Episode: 426 \t\t Reward: -114.04049999999974\n",
      "Episode: 427 \t\t Reward: -112.9519999999997\n",
      "Episode: 428 \t\t Reward: -112.4324999999997\n",
      "Episode: 429 \t\t Reward: -107.1254999999997\n",
      "Episode: 430 \t\t Reward: -107.02049999999971\n",
      "Episode: 431 \t\t Reward: -112.06849999999972\n",
      "Episode: 432 \t\t Reward: -111.47799999999972\n",
      "Episode: 433 \t\t Reward: -111.6329999999997\n",
      "Episode: 434 \t\t Reward: -110.7004999999997\n",
      "Episode: 435 \t\t Reward: -111.17649999999972\n",
      "Episode: 436 \t\t Reward: -109.87849999999968\n",
      "Episode: 437 \t\t Reward: -109.58449999999968\n",
      "Episode: 438 \t\t Reward: -109.83149999999966\n",
      "Episode: 439 \t\t Reward: -110.31949999999969\n",
      "Episode: 440 \t\t Reward: -110.24249999999971\n",
      "Episode: 441 \t\t Reward: -109.09299999999968\n",
      "Episode: 442 \t\t Reward: -114.4319999999997\n",
      "Episode: 443 \t\t Reward: -114.64549999999969\n",
      "Episode: 444 \t\t Reward: -114.86499999999971\n",
      "Episode: 445 \t\t Reward: -115.13999999999969\n",
      "Episode: 446 \t\t Reward: -114.99899999999968\n",
      "Episode: 447 \t\t Reward: -119.80549999999968\n",
      "Episode: 448 \t\t Reward: -124.5644999999997\n",
      "Episode: 449 \t\t Reward: -123.92099999999965\n",
      "Episode: 450 \t\t Reward: -125.30549999999968\n",
      "Episode: 451 \t\t Reward: -126.12049999999972\n",
      "Episode: 452 \t\t Reward: -125.61599999999972\n",
      "Episode: 453 \t\t Reward: -126.07599999999971\n",
      "Episode: 454 \t\t Reward: -121.97449999999975\n",
      "Episode: 455 \t\t Reward: -122.20149999999973\n",
      "Episode: 456 \t\t Reward: -121.14849999999973\n",
      "Episode: 457 \t\t Reward: -121.57549999999972\n",
      "Episode: 458 \t\t Reward: -122.01349999999974\n",
      "Episode: 459 \t\t Reward: -121.42499999999971\n",
      "Episode: 460 \t\t Reward: -120.90299999999972\n",
      "Episode: 461 \t\t Reward: -120.36999999999969\n",
      "Episode: 462 \t\t Reward: -124.3859999999997\n",
      "Episode: 463 \t\t Reward: -123.61299999999969\n",
      "Episode: 464 \t\t Reward: -123.66499999999965\n",
      "Episode: 465 \t\t Reward: -118.70499999999967\n",
      "Episode: 466 \t\t Reward: -119.05299999999967\n",
      "Episode: 467 \t\t Reward: -118.99399999999967\n",
      "Episode: 468 \t\t Reward: -121.41549999999967\n",
      "Episode: 469 \t\t Reward: -125.8794999999997\n",
      "Episode: 470 \t\t Reward: -126.2329999999997\n",
      "Episode: 471 \t\t Reward: -121.06349999999973\n",
      "Episode: 472 \t\t Reward: -125.3274999999997\n",
      "Episode: 473 \t\t Reward: -126.24449999999972\n",
      "Episode: 474 \t\t Reward: -126.37699999999974\n",
      "Episode: 475 \t\t Reward: -131.7164999999997\n",
      "Episode: 476 \t\t Reward: -132.18449999999973\n",
      "Episode: 477 \t\t Reward: -132.86749999999975\n",
      "Episode: 478 \t\t Reward: -132.58399999999972\n",
      "Episode: 479 \t\t Reward: -133.06599999999972\n",
      "Episode: 480 \t\t Reward: -133.20849999999976\n",
      "Episode: 481 \t\t Reward: -128.03799999999973\n",
      "Episode: 482 \t\t Reward: -128.86949999999973\n",
      "Episode: 483 \t\t Reward: -135.15699999999975\n",
      "Episode: 484 \t\t Reward: -135.31899999999973\n",
      "Episode: 485 \t\t Reward: -134.52349999999976\n",
      "Episode: 486 \t\t Reward: -129.74449999999976\n",
      "Episode: 487 \t\t Reward: -129.12699999999973\n",
      "Episode: 488 \t\t Reward: -129.15299999999976\n",
      "Episode: 489 \t\t Reward: -129.5229999999998\n",
      "Episode: 490 \t\t Reward: -128.41999999999973\n",
      "Episode: 491 \t\t Reward: -134.07349999999974\n",
      "Episode: 492 \t\t Reward: -133.98799999999974\n",
      "Episode: 493 \t\t Reward: -134.35049999999976\n",
      "Episode: 494 \t\t Reward: -134.31999999999974\n",
      "Episode: 495 \t\t Reward: -134.20849999999973\n",
      "Episode: 496 \t\t Reward: -134.41949999999974\n",
      "Episode: 497 \t\t Reward: -134.76549999999975\n",
      "Episode: 498 \t\t Reward: -134.50349999999972\n",
      "Episode: 499 \t\t Reward: -134.88399999999973\n",
      "Episode: 500 \t\t Reward: -134.38799999999975\n",
      "Episode: 501 \t\t Reward: -139.40599999999975\n",
      "Episode: 502 \t\t Reward: -138.93249999999975\n",
      "Episode: 503 \t\t Reward: -140.07549999999975\n",
      "Episode: 504 \t\t Reward: -134.67949999999976\n",
      "Episode: 505 \t\t Reward: -134.19399999999976\n",
      "Episode: 506 \t\t Reward: -133.07599999999974\n",
      "Episode: 507 \t\t Reward: -133.49599999999975\n",
      "Episode: 508 \t\t Reward: -133.18549999999973\n",
      "Episode: 509 \t\t Reward: -132.97449999999972\n",
      "Episode: 510 \t\t Reward: -133.40599999999975\n",
      "Episode: 511 \t\t Reward: -133.26299999999975\n",
      "Episode: 512 \t\t Reward: -132.5514999999997\n",
      "Episode: 513 \t\t Reward: -127.20199999999973\n",
      "Episode: 514 \t\t Reward: -131.51249999999973\n",
      "Episode: 515 \t\t Reward: -131.3114999999997\n",
      "Episode: 516 \t\t Reward: -131.97499999999974\n",
      "Episode: 517 \t\t Reward: -137.35849999999974\n",
      "Episode: 518 \t\t Reward: -132.20949999999974\n",
      "Episode: 519 \t\t Reward: -131.8404999999997\n",
      "Episode: 520 \t\t Reward: -134.21849999999978\n",
      "Episode: 521 \t\t Reward: -135.15499999999975\n",
      "Episode: 522 \t\t Reward: -135.20599999999976\n",
      "Episode: 523 \t\t Reward: -134.50949999999972\n",
      "Episode: 524 \t\t Reward: -139.58699999999973\n",
      "Episode: 525 \t\t Reward: -134.99499999999972\n",
      "Episode: 526 \t\t Reward: -135.11699999999973\n",
      "Episode: 527 \t\t Reward: -135.82149999999976\n",
      "Episode: 528 \t\t Reward: -135.46799999999973\n",
      "Episode: 529 \t\t Reward: -135.62149999999974\n",
      "Episode: 530 \t\t Reward: -135.45399999999975\n",
      "Episode: 531 \t\t Reward: -129.70399999999975\n",
      "Episode: 532 \t\t Reward: -130.37549999999976\n",
      "Episode: 533 \t\t Reward: -134.8899999999998\n",
      "Episode: 534 \t\t Reward: -135.21699999999976\n",
      "Episode: 535 \t\t Reward: -129.94499999999977\n",
      "Episode: 536 \t\t Reward: -131.1739999999998\n",
      "Episode: 537 \t\t Reward: -131.0374999999998\n",
      "Episode: 538 \t\t Reward: -130.06799999999978\n",
      "Episode: 539 \t\t Reward: -130.47199999999975\n",
      "Episode: 540 \t\t Reward: -130.15249999999978\n",
      "Episode: 541 \t\t Reward: -130.69549999999978\n",
      "Episode: 542 \t\t Reward: -125.3874999999998\n",
      "Episode: 543 \t\t Reward: -124.6089999999998\n",
      "Episode: 544 \t\t Reward: -125.09349999999978\n",
      "Episode: 545 \t\t Reward: -125.04349999999978\n",
      "Episode: 546 \t\t Reward: -124.69449999999978\n",
      "Episode: 547 \t\t Reward: -119.86449999999975\n",
      "Episode: 548 \t\t Reward: -114.56299999999978\n",
      "Episode: 549 \t\t Reward: -115.13699999999977\n",
      "Episode: 550 \t\t Reward: -114.74549999999978\n",
      "Episode: 551 \t\t Reward: -119.84199999999977\n",
      "Episode: 552 \t\t Reward: -119.84199999999979\n",
      "Episode: 553 \t\t Reward: -119.66149999999978\n",
      "Episode: 554 \t\t Reward: -119.10349999999977\n",
      "Episode: 555 \t\t Reward: -118.98549999999977\n",
      "Episode: 556 \t\t Reward: -124.5939999999998\n",
      "Episode: 557 \t\t Reward: -124.4199999999998\n",
      "Episode: 558 \t\t Reward: -123.4894999999998\n",
      "Episode: 559 \t\t Reward: -128.4739999999998\n",
      "Episode: 560 \t\t Reward: -129.14049999999978\n",
      "Episode: 561 \t\t Reward: -129.2949999999998\n",
      "Episode: 562 \t\t Reward: -122.59699999999977\n",
      "Episode: 563 \t\t Reward: -121.87449999999973\n",
      "Episode: 564 \t\t Reward: -122.37799999999977\n",
      "Episode: 565 \t\t Reward: -122.03099999999975\n",
      "Episode: 566 \t\t Reward: -116.55749999999975\n",
      "Episode: 567 \t\t Reward: -116.54699999999974\n",
      "Episode: 568 \t\t Reward: -116.11149999999974\n",
      "Episode: 569 \t\t Reward: -116.44799999999974\n",
      "Episode: 570 \t\t Reward: -116.86849999999976\n",
      "Episode: 571 \t\t Reward: -116.98449999999976\n",
      "Episode: 572 \t\t Reward: -112.58899999999976\n",
      "Episode: 573 \t\t Reward: -111.94249999999977\n",
      "Episode: 574 \t\t Reward: -111.20849999999973\n",
      "Episode: 575 \t\t Reward: -105.36049999999976\n",
      "Episode: 576 \t\t Reward: -100.04249999999973\n",
      "Episode: 577 \t\t Reward: -100.56699999999974\n",
      "Episode: 578 \t\t Reward: -101.42649999999975\n",
      "Episode: 579 \t\t Reward: -106.24499999999975\n",
      "Episode: 580 \t\t Reward: -101.28749999999977\n",
      "Episode: 581 \t\t Reward: -101.57649999999975\n",
      "Episode: 582 \t\t Reward: -100.80349999999974\n",
      "Episode: 583 \t\t Reward: -95.41899999999976\n",
      "Episode: 584 \t\t Reward: -95.01099999999975\n",
      "Episode: 585 \t\t Reward: -101.12249999999977\n",
      "Episode: 586 \t\t Reward: -101.29649999999978\n",
      "Episode: 587 \t\t Reward: -101.71849999999978\n",
      "Episode: 588 \t\t Reward: -101.7699999999998\n",
      "Episode: 589 \t\t Reward: -99.48299999999976\n",
      "Episode: 590 \t\t Reward: -100.41149999999976\n",
      "Episode: 591 \t\t Reward: -95.19799999999975\n",
      "Episode: 592 \t\t Reward: -94.52549999999974\n",
      "Episode: 593 \t\t Reward: -94.73149999999973\n",
      "Episode: 594 \t\t Reward: -88.78699999999972\n",
      "Episode: 595 \t\t Reward: -88.47049999999972\n",
      "Episode: 596 \t\t Reward: -89.06399999999977\n",
      "Episode: 597 \t\t Reward: -88.84449999999977\n",
      "Episode: 598 \t\t Reward: -89.27249999999977\n",
      "Episode: 599 \t\t Reward: -88.82549999999975\n",
      "Episode: 600 \t\t Reward: -89.25899999999974\n",
      "Episode: 601 \t\t Reward: -88.66249999999972\n",
      "Episode: 602 \t\t Reward: -88.98549999999973\n",
      "Episode: 603 \t\t Reward: -88.56549999999974\n",
      "Episode: 604 \t\t Reward: -93.20149999999974\n",
      "Episode: 605 \t\t Reward: -98.68649999999974\n",
      "Episode: 606 \t\t Reward: -94.70499999999976\n",
      "Episode: 607 \t\t Reward: -94.81449999999977\n",
      "Episode: 608 \t\t Reward: -94.26349999999974\n",
      "Episode: 609 \t\t Reward: -94.38899999999975\n",
      "Episode: 610 \t\t Reward: -94.25549999999974\n",
      "Episode: 611 \t\t Reward: -95.24249999999975\n",
      "Episode: 612 \t\t Reward: -93.84249999999973\n",
      "Episode: 613 \t\t Reward: -93.46999999999971\n",
      "Episode: 614 \t\t Reward: -89.53799999999974\n",
      "Episode: 615 \t\t Reward: -94.89249999999973\n",
      "Episode: 616 \t\t Reward: -94.89099999999974\n",
      "Episode: 617 \t\t Reward: -89.57949999999974\n",
      "Episode: 618 \t\t Reward: -89.13449999999973\n",
      "Episode: 619 \t\t Reward: -89.31049999999973\n",
      "Episode: 620 \t\t Reward: -89.00049999999973\n",
      "Episode: 621 \t\t Reward: -94.12899999999973\n",
      "Episode: 622 \t\t Reward: -94.11849999999976\n",
      "Episode: 623 \t\t Reward: -95.05349999999979\n",
      "Episode: 624 \t\t Reward: -95.42649999999975\n",
      "Episode: 625 \t\t Reward: -95.29899999999977\n",
      "Episode: 626 \t\t Reward: -95.21599999999972\n",
      "Episode: 627 \t\t Reward: -95.80049999999976\n",
      "Episode: 628 \t\t Reward: -96.40849999999979\n",
      "Episode: 629 \t\t Reward: -101.4399999999998\n",
      "Episode: 630 \t\t Reward: -101.42949999999979\n",
      "Episode: 631 \t\t Reward: -101.87999999999982\n",
      "Episode: 632 \t\t Reward: -106.8174999999998\n",
      "Episode: 633 \t\t Reward: -102.3574999999998\n",
      "Episode: 634 \t\t Reward: -102.2319999999998\n",
      "Episode: 635 \t\t Reward: -102.19849999999978\n",
      "Episode: 636 \t\t Reward: -107.12549999999978\n",
      "Episode: 637 \t\t Reward: -106.9594999999998\n",
      "Episode: 638 \t\t Reward: -107.4159999999998\n",
      "Episode: 639 \t\t Reward: -107.08449999999979\n",
      "Episode: 640 \t\t Reward: -112.1864999999998\n",
      "Episode: 641 \t\t Reward: -112.4974999999998\n",
      "Episode: 642 \t\t Reward: -117.45399999999981\n",
      "Episode: 643 \t\t Reward: -118.21149999999982\n",
      "Episode: 644 \t\t Reward: -118.25599999999982\n",
      "Episode: 645 \t\t Reward: -118.84249999999982\n",
      "Episode: 646 \t\t Reward: -118.95599999999982\n",
      "Episode: 647 \t\t Reward: -124.46149999999982\n",
      "Episode: 648 \t\t Reward: -124.27849999999981\n",
      "Episode: 649 \t\t Reward: -124.16849999999981\n",
      "Episode: 650 \t\t Reward: -123.7964999999998\n",
      "Episode: 651 \t\t Reward: -118.8854999999998\n",
      "Episode: 652 \t\t Reward: -119.6479999999998\n",
      "Episode: 653 \t\t Reward: -118.99749999999982\n",
      "Episode: 654 \t\t Reward: -119.88299999999983\n",
      "Episode: 655 \t\t Reward: -119.41599999999983\n",
      "Episode: 656 \t\t Reward: -114.55249999999982\n",
      "Episode: 657 \t\t Reward: -114.63249999999984\n",
      "Episode: 658 \t\t Reward: -115.59449999999987\n",
      "Episode: 659 \t\t Reward: -110.85399999999986\n",
      "Episode: 660 \t\t Reward: -115.42099999999984\n",
      "Episode: 661 \t\t Reward: -120.45899999999985\n",
      "Episode: 662 \t\t Reward: -128.19949999999986\n",
      "Episode: 663 \t\t Reward: -128.73149999999987\n",
      "Episode: 664 \t\t Reward: -128.77899999999988\n",
      "Episode: 665 \t\t Reward: -128.73499999999987\n",
      "Episode: 666 \t\t Reward: -128.48599999999988\n",
      "Episode: 667 \t\t Reward: -128.3889999999999\n",
      "Episode: 668 \t\t Reward: -128.13149999999987\n",
      "Episode: 669 \t\t Reward: -123.12749999999986\n",
      "Episode: 670 \t\t Reward: -122.61699999999985\n",
      "Episode: 671 \t\t Reward: -122.16849999999985\n",
      "Episode: 672 \t\t Reward: -122.45849999999986\n",
      "Episode: 673 \t\t Reward: -123.08499999999985\n",
      "Episode: 674 \t\t Reward: -123.76199999999989\n",
      "Episode: 675 \t\t Reward: -129.83299999999986\n",
      "Episode: 676 \t\t Reward: -130.14799999999988\n",
      "Episode: 677 \t\t Reward: -127.83799999999985\n",
      "Episode: 678 \t\t Reward: -127.06899999999983\n",
      "Episode: 679 \t\t Reward: -121.73299999999983\n",
      "Episode: 680 \t\t Reward: -121.47399999999982\n",
      "Episode: 681 \t\t Reward: -120.88199999999982\n",
      "Episode: 682 \t\t Reward: -121.42849999999984\n",
      "Episode: 683 \t\t Reward: -121.01149999999981\n",
      "Episode: 684 \t\t Reward: -118.89649999999978\n",
      "Episode: 685 \t\t Reward: -113.39999999999975\n",
      "Episode: 686 \t\t Reward: -113.46649999999975\n",
      "Episode: 687 \t\t Reward: -118.37099999999977\n",
      "Episode: 688 \t\t Reward: -118.40299999999978\n",
      "Episode: 689 \t\t Reward: -125.47799999999981\n",
      "Episode: 690 \t\t Reward: -125.4169999999998\n",
      "Episode: 691 \t\t Reward: -130.6729999999998\n",
      "Episode: 692 \t\t Reward: -131.28049999999982\n",
      "Episode: 693 \t\t Reward: -136.03649999999982\n",
      "Episode: 694 \t\t Reward: -136.64999999999984\n",
      "Episode: 695 \t\t Reward: -136.48299999999983\n",
      "Episode: 696 \t\t Reward: -135.6119999999998\n",
      "Episode: 697 \t\t Reward: -135.8959999999998\n",
      "Episode: 698 \t\t Reward: -135.91399999999982\n",
      "Episode: 699 \t\t Reward: -135.9804999999998\n",
      "Episode: 700 \t\t Reward: -136.09649999999982\n",
      "Episode: 701 \t\t Reward: -130.8454999999998\n",
      "Episode: 702 \t\t Reward: -136.1994999999998\n",
      "Episode: 703 \t\t Reward: -135.82649999999978\n",
      "Episode: 704 \t\t Reward: -131.2634999999998\n",
      "Episode: 705 \t\t Reward: -125.61449999999981\n",
      "Episode: 706 \t\t Reward: -125.55449999999982\n",
      "Episode: 707 \t\t Reward: -125.80799999999981\n",
      "Episode: 708 \t\t Reward: -126.53549999999984\n",
      "Episode: 709 \t\t Reward: -126.59749999999984\n",
      "Episode: 710 \t\t Reward: -126.55649999999983\n",
      "Episode: 711 \t\t Reward: -125.92199999999981\n",
      "Episode: 712 \t\t Reward: -128.35299999999987\n",
      "Episode: 713 \t\t Reward: -128.43099999999987\n",
      "Episode: 714 \t\t Reward: -132.93949999999987\n",
      "Episode: 715 \t\t Reward: -127.20699999999987\n",
      "Episode: 716 \t\t Reward: -126.84699999999987\n",
      "Episode: 717 \t\t Reward: -126.45499999999984\n",
      "Episode: 718 \t\t Reward: -126.89349999999985\n",
      "Episode: 719 \t\t Reward: -127.04599999999986\n",
      "Episode: 720 \t\t Reward: -127.51949999999984\n",
      "Episode: 721 \t\t Reward: -121.99299999999987\n",
      "Episode: 722 \t\t Reward: -126.69199999999987\n",
      "Episode: 723 \t\t Reward: -126.74749999999986\n",
      "Episode: 724 \t\t Reward: -126.25649999999985\n",
      "Episode: 725 \t\t Reward: -131.86649999999986\n",
      "Episode: 726 \t\t Reward: -131.5719999999999\n",
      "Episode: 727 \t\t Reward: -136.2679999999999\n",
      "Episode: 728 \t\t Reward: -135.83399999999986\n",
      "Episode: 729 \t\t Reward: -130.49049999999986\n",
      "Episode: 730 \t\t Reward: -130.75399999999985\n",
      "Episode: 731 \t\t Reward: -130.41049999999987\n",
      "Episode: 732 \t\t Reward: -124.88949999999983\n",
      "Episode: 733 \t\t Reward: -128.88449999999983\n",
      "Episode: 734 \t\t Reward: -129.46649999999983\n",
      "Episode: 735 \t\t Reward: -134.0774999999998\n",
      "Episode: 736 \t\t Reward: -129.08449999999982\n",
      "Episode: 737 \t\t Reward: -129.08999999999978\n",
      "Episode: 738 \t\t Reward: -134.9709999999998\n",
      "Episode: 739 \t\t Reward: -135.2689999999998\n",
      "Episode: 740 \t\t Reward: -130.78799999999984\n",
      "Episode: 741 \t\t Reward: -135.47349999999983\n",
      "Episode: 742 \t\t Reward: -130.32949999999985\n",
      "Episode: 743 \t\t Reward: -129.0739999999998\n",
      "Episode: 744 \t\t Reward: -128.5619999999998\n",
      "Episode: 745 \t\t Reward: -128.32049999999978\n",
      "Episode: 746 \t\t Reward: -128.8014999999998\n",
      "Episode: 747 \t\t Reward: -122.45199999999979\n",
      "Episode: 748 \t\t Reward: -127.9134999999998\n",
      "Episode: 749 \t\t Reward: -128.06899999999982\n",
      "Episode: 750 \t\t Reward: -133.4999999999998\n",
      "Episode: 751 \t\t Reward: -133.0549999999998\n",
      "Episode: 752 \t\t Reward: -137.9464999999998\n",
      "Episode: 753 \t\t Reward: -138.7694999999998\n",
      "Episode: 754 \t\t Reward: -138.69949999999977\n",
      "Episode: 755 \t\t Reward: -139.0164999999998\n",
      "Episode: 756 \t\t Reward: -139.31349999999978\n",
      "Episode: 757 \t\t Reward: -138.07499999999976\n",
      "Episode: 758 \t\t Reward: -136.83799999999974\n",
      "Episode: 759 \t\t Reward: -136.90649999999974\n",
      "Episode: 760 \t\t Reward: -137.20999999999975\n",
      "Episode: 761 \t\t Reward: -132.55949999999976\n",
      "Episode: 762 \t\t Reward: -127.34549999999976\n",
      "Episode: 763 \t\t Reward: -127.66199999999975\n",
      "Episode: 764 \t\t Reward: -132.99099999999973\n",
      "Episode: 765 \t\t Reward: -131.22199999999972\n",
      "Episode: 766 \t\t Reward: -132.05149999999975\n",
      "Episode: 767 \t\t Reward: -131.70599999999973\n",
      "Episode: 768 \t\t Reward: -132.42199999999974\n",
      "Episode: 769 \t\t Reward: -132.73949999999977\n",
      "Episode: 770 \t\t Reward: -133.26749999999976\n",
      "Episode: 771 \t\t Reward: -133.40599999999978\n",
      "Episode: 772 \t\t Reward: -132.51849999999976\n",
      "Episode: 773 \t\t Reward: -132.16349999999974\n",
      "Episode: 774 \t\t Reward: -131.51749999999973\n",
      "Episode: 775 \t\t Reward: -131.55799999999974\n",
      "Episode: 776 \t\t Reward: -130.5514999999997\n",
      "Episode: 777 \t\t Reward: -133.18449999999973\n",
      "Episode: 778 \t\t Reward: -131.5534999999997\n",
      "Episode: 779 \t\t Reward: -130.95799999999969\n",
      "Episode: 780 \t\t Reward: -128.98499999999964\n",
      "Episode: 781 \t\t Reward: -127.13899999999963\n",
      "Episode: 782 \t\t Reward: -127.19349999999964\n",
      "Episode: 783 \t\t Reward: -127.32299999999965\n",
      "Episode: 784 \t\t Reward: -134.8374999999997\n",
      "Episode: 785 \t\t Reward: -135.15999999999968\n",
      "Episode: 786 \t\t Reward: -133.67749999999967\n",
      "Episode: 787 \t\t Reward: -128.86599999999964\n",
      "Episode: 788 \t\t Reward: -128.38549999999964\n",
      "Episode: 789 \t\t Reward: -123.60099999999964\n",
      "Episode: 790 \t\t Reward: -123.49899999999963\n",
      "Episode: 791 \t\t Reward: -118.26749999999966\n",
      "Episode: 792 \t\t Reward: -118.03649999999966\n",
      "Episode: 793 \t\t Reward: -112.74349999999964\n",
      "Episode: 794 \t\t Reward: -112.82849999999966\n",
      "Episode: 795 \t\t Reward: -112.68899999999964\n",
      "Episode: 796 \t\t Reward: -113.53299999999966\n",
      "Episode: 797 \t\t Reward: -118.06949999999966\n",
      "Episode: 798 \t\t Reward: -117.73149999999967\n",
      "Episode: 799 \t\t Reward: -118.36099999999966\n",
      "Episode: 800 \t\t Reward: -117.72699999999966\n",
      "Episode: 801 \t\t Reward: -117.93399999999966\n",
      "Episode: 802 \t\t Reward: -111.61649999999963\n",
      "Episode: 803 \t\t Reward: -116.29799999999963\n",
      "Episode: 804 \t\t Reward: -121.28499999999964\n",
      "Episode: 805 \t\t Reward: -121.80649999999963\n",
      "Episode: 806 \t\t Reward: -121.76999999999964\n",
      "Episode: 807 \t\t Reward: -121.41449999999965\n",
      "Episode: 808 \t\t Reward: -120.4019999999996\n",
      "Episode: 809 \t\t Reward: -125.5589999999996\n",
      "Episode: 810 \t\t Reward: -126.1799999999996\n",
      "Episode: 811 \t\t Reward: -126.2759999999996\n",
      "Episode: 812 \t\t Reward: -125.8349999999996\n",
      "Episode: 813 \t\t Reward: -125.72749999999962\n",
      "Episode: 814 \t\t Reward: -121.10149999999963\n",
      "Episode: 815 \t\t Reward: -120.53299999999959\n",
      "Episode: 816 \t\t Reward: -120.4874999999996\n",
      "Episode: 817 \t\t Reward: -120.78899999999959\n",
      "Episode: 818 \t\t Reward: -120.35049999999961\n",
      "Episode: 819 \t\t Reward: -118.33349999999959\n",
      "Episode: 820 \t\t Reward: -117.92649999999959\n",
      "Episode: 821 \t\t Reward: -118.26749999999959\n",
      "Episode: 822 \t\t Reward: -113.37199999999959\n",
      "Episode: 823 \t\t Reward: -113.63049999999959\n",
      "Episode: 824 \t\t Reward: -113.77849999999957\n",
      "Episode: 825 \t\t Reward: -108.38949999999959\n",
      "Episode: 826 \t\t Reward: -108.27099999999959\n",
      "Episode: 827 \t\t Reward: -103.67549999999959\n",
      "Episode: 828 \t\t Reward: -104.29549999999959\n",
      "Episode: 829 \t\t Reward: -104.80649999999962\n",
      "Episode: 830 \t\t Reward: -104.6509999999996\n",
      "Episode: 831 \t\t Reward: -105.2719999999996\n",
      "Episode: 832 \t\t Reward: -105.72899999999962\n",
      "Episode: 833 \t\t Reward: -101.24899999999965\n",
      "Episode: 834 \t\t Reward: -101.25299999999964\n",
      "Episode: 835 \t\t Reward: -96.52299999999964\n",
      "Episode: 836 \t\t Reward: -96.12249999999966\n",
      "Episode: 837 \t\t Reward: -96.40299999999965\n",
      "Episode: 838 \t\t Reward: -90.05399999999963\n",
      "Episode: 839 \t\t Reward: -95.53399999999964\n",
      "Episode: 840 \t\t Reward: -95.33649999999962\n",
      "Episode: 841 \t\t Reward: -90.26699999999961\n",
      "Episode: 842 \t\t Reward: -90.5234999999996\n",
      "Episode: 843 \t\t Reward: -89.4774999999996\n",
      "Episode: 844 \t\t Reward: -94.55449999999959\n",
      "Episode: 845 \t\t Reward: -94.06799999999957\n",
      "Episode: 846 \t\t Reward: -93.97999999999958\n",
      "Episode: 847 \t\t Reward: -94.4639999999996\n",
      "Episode: 848 \t\t Reward: -89.19949999999959\n",
      "Episode: 849 \t\t Reward: -89.06049999999959\n",
      "Episode: 850 \t\t Reward: -83.9634999999996\n",
      "Episode: 851 \t\t Reward: -83.56799999999959\n",
      "Episode: 852 \t\t Reward: -83.7024999999996\n",
      "Episode: 853 \t\t Reward: -82.34749999999958\n",
      "Episode: 854 \t\t Reward: -81.69299999999957\n",
      "Episode: 855 \t\t Reward: -81.75149999999957\n",
      "Episode: 856 \t\t Reward: -81.76049999999955\n",
      "Episode: 857 \t\t Reward: -83.02999999999957\n",
      "Episode: 858 \t\t Reward: -83.9154999999996\n",
      "Episode: 859 \t\t Reward: -83.23799999999959\n",
      "Episode: 860 \t\t Reward: -83.46799999999959\n",
      "Episode: 861 \t\t Reward: -88.2514999999996\n",
      "Episode: 862 \t\t Reward: -88.32949999999961\n",
      "Episode: 863 \t\t Reward: -88.9309999999996\n",
      "Episode: 864 \t\t Reward: -83.7109999999996\n",
      "Episode: 865 \t\t Reward: -85.85199999999965\n",
      "Episode: 866 \t\t Reward: -85.10999999999963\n",
      "Episode: 867 \t\t Reward: -85.93849999999964\n",
      "Episode: 868 \t\t Reward: -85.81299999999963\n",
      "Episode: 869 \t\t Reward: -84.91649999999962\n",
      "Episode: 870 \t\t Reward: -84.35999999999962\n",
      "Episode: 871 \t\t Reward: -83.51449999999957\n",
      "Episode: 872 \t\t Reward: -83.15299999999958\n",
      "Episode: 873 \t\t Reward: -83.37899999999956\n",
      "Episode: 874 \t\t Reward: -84.0509999999996\n",
      "Episode: 875 \t\t Reward: -78.42899999999958\n",
      "Episode: 876 \t\t Reward: -78.9574999999996\n",
      "Episode: 877 \t\t Reward: -78.3564999999996\n",
      "Episode: 878 \t\t Reward: -80.07499999999965\n",
      "Episode: 879 \t\t Reward: -80.53749999999965\n",
      "Episode: 880 \t\t Reward: -82.6744999999997\n",
      "Episode: 881 \t\t Reward: -84.1659999999997\n",
      "Episode: 882 \t\t Reward: -88.75299999999969\n",
      "Episode: 883 \t\t Reward: -94.5224999999997\n",
      "Episode: 884 \t\t Reward: -89.4169999999997\n",
      "Episode: 885 \t\t Reward: -89.6114999999997\n",
      "Episode: 886 \t\t Reward: -90.92099999999974\n",
      "Episode: 887 \t\t Reward: -90.73749999999974\n",
      "Episode: 888 \t\t Reward: -91.63549999999974\n",
      "Episode: 889 \t\t Reward: -90.94949999999973\n",
      "Episode: 890 \t\t Reward: -90.80249999999972\n",
      "Episode: 891 \t\t Reward: -90.15349999999971\n",
      "Episode: 892 \t\t Reward: -90.38149999999969\n",
      "Episode: 893 \t\t Reward: -95.2839999999997\n",
      "Episode: 894 \t\t Reward: -94.33349999999967\n",
      "Episode: 895 \t\t Reward: -95.01549999999972\n",
      "Episode: 896 \t\t Reward: -94.89949999999972\n",
      "Episode: 897 \t\t Reward: -89.78199999999971\n",
      "Episode: 898 \t\t Reward: -89.72549999999973\n",
      "Episode: 899 \t\t Reward: -93.8649999999997\n",
      "Episode: 900 \t\t Reward: -93.96799999999968\n",
      "Episode: 901 \t\t Reward: -94.46699999999969\n",
      "Episode: 902 \t\t Reward: -94.72399999999966\n",
      "Episode: 903 \t\t Reward: -89.9184999999997\n",
      "Episode: 904 \t\t Reward: -84.33349999999969\n",
      "Episode: 905 \t\t Reward: -84.50299999999969\n",
      "Episode: 906 \t\t Reward: -88.76549999999969\n",
      "Episode: 907 \t\t Reward: -89.13699999999967\n",
      "Episode: 908 \t\t Reward: -95.56699999999972\n",
      "Episode: 909 \t\t Reward: -90.56099999999971\n",
      "Episode: 910 \t\t Reward: -89.97199999999971\n",
      "Episode: 911 \t\t Reward: -89.87649999999971\n",
      "Episode: 912 \t\t Reward: -90.3839999999997\n",
      "Episode: 913 \t\t Reward: -95.55799999999971\n",
      "Episode: 914 \t\t Reward: -95.60549999999968\n",
      "Episode: 915 \t\t Reward: -95.97949999999972\n",
      "Episode: 916 \t\t Reward: -96.55099999999972\n",
      "Episode: 917 \t\t Reward: -96.46799999999972\n",
      "Episode: 918 \t\t Reward: -101.95799999999971\n",
      "Episode: 919 \t\t Reward: -109.60049999999976\n",
      "Episode: 920 \t\t Reward: -110.05899999999974\n",
      "Episode: 921 \t\t Reward: -110.07649999999973\n",
      "Episode: 922 \t\t Reward: -107.8559999999997\n",
      "Episode: 923 \t\t Reward: -107.7124999999997\n",
      "Episode: 924 \t\t Reward: -102.45649999999969\n",
      "Episode: 925 \t\t Reward: -102.1514999999997\n",
      "Episode: 926 \t\t Reward: -102.62849999999972\n",
      "Episode: 927 \t\t Reward: -102.24749999999969\n",
      "Episode: 928 \t\t Reward: -102.2934999999997\n",
      "Episode: 929 \t\t Reward: -101.4524999999997\n",
      "Episode: 930 \t\t Reward: -101.59499999999971\n",
      "Episode: 931 \t\t Reward: -101.55099999999972\n",
      "Episode: 932 \t\t Reward: -101.79199999999972\n",
      "Episode: 933 \t\t Reward: -102.19049999999972\n",
      "Episode: 934 \t\t Reward: -102.23349999999971\n",
      "Episode: 935 \t\t Reward: -107.81749999999971\n",
      "Episode: 936 \t\t Reward: -106.85099999999966\n",
      "Episode: 937 \t\t Reward: -106.75349999999969\n",
      "Episode: 938 \t\t Reward: -107.89349999999969\n",
      "Episode: 939 \t\t Reward: -102.8344999999997\n",
      "Episode: 940 \t\t Reward: -102.4229999999997\n",
      "Episode: 941 \t\t Reward: -102.9529999999997\n",
      "Episode: 942 \t\t Reward: -102.7149999999997\n",
      "Episode: 943 \t\t Reward: -104.47149999999974\n",
      "Episode: 944 \t\t Reward: -99.04749999999972\n",
      "Episode: 945 \t\t Reward: -99.52649999999973\n",
      "Episode: 946 \t\t Reward: -99.5364999999997\n",
      "Episode: 947 \t\t Reward: -104.86149999999972\n",
      "Episode: 948 \t\t Reward: -110.13749999999973\n",
      "Episode: 949 \t\t Reward: -110.04349999999972\n",
      "Episode: 950 \t\t Reward: -109.58249999999973\n",
      "Episode: 951 \t\t Reward: -110.59999999999974\n",
      "Episode: 952 \t\t Reward: -105.48949999999975\n",
      "Episode: 953 \t\t Reward: -110.94599999999977\n",
      "Episode: 954 \t\t Reward: -110.93599999999977\n",
      "Episode: 955 \t\t Reward: -109.72099999999973\n",
      "Episode: 956 \t\t Reward: -108.83199999999974\n",
      "Episode: 957 \t\t Reward: -108.41649999999974\n",
      "Episode: 958 \t\t Reward: -108.40399999999974\n",
      "Episode: 959 \t\t Reward: -108.52949999999973\n",
      "Episode: 960 \t\t Reward: -102.76749999999973\n",
      "Episode: 961 \t\t Reward: -97.91949999999972\n",
      "Episode: 962 \t\t Reward: -97.71949999999971\n",
      "Episode: 963 \t\t Reward: -97.57099999999971\n",
      "Episode: 964 \t\t Reward: -97.11149999999972\n",
      "Episode: 965 \t\t Reward: -96.88949999999971\n",
      "Episode: 966 \t\t Reward: -97.3579999999997\n",
      "Episode: 967 \t\t Reward: -97.16799999999972\n",
      "Episode: 968 \t\t Reward: -97.31749999999973\n",
      "Episode: 969 \t\t Reward: -98.18849999999973\n",
      "Episode: 970 \t\t Reward: -98.40099999999975\n",
      "Episode: 971 \t\t Reward: -104.67899999999977\n",
      "Episode: 972 \t\t Reward: -105.32199999999978\n",
      "Episode: 973 \t\t Reward: -104.13199999999975\n",
      "Episode: 974 \t\t Reward: -104.08849999999975\n",
      "Episode: 975 \t\t Reward: -104.71999999999973\n",
      "Episode: 976 \t\t Reward: -105.07799999999973\n",
      "Episode: 977 \t\t Reward: -105.45799999999976\n",
      "Episode: 978 \t\t Reward: -105.36849999999976\n",
      "Episode: 979 \t\t Reward: -104.86249999999974\n",
      "Episode: 980 \t\t Reward: -110.45849999999975\n",
      "Episode: 981 \t\t Reward: -111.28749999999974\n",
      "Episode: 982 \t\t Reward: -106.67549999999976\n",
      "Episode: 983 \t\t Reward: -100.94249999999974\n",
      "Episode: 984 \t\t Reward: -100.22349999999973\n",
      "Episode: 985 \t\t Reward: -99.41599999999973\n",
      "Episode: 986 \t\t Reward: -99.5994999999997\n",
      "Episode: 987 \t\t Reward: -99.8399999999997\n",
      "Episode: 988 \t\t Reward: -99.3269999999997\n",
      "Episode: 989 \t\t Reward: -99.85399999999973\n",
      "Episode: 990 \t\t Reward: -99.47499999999974\n",
      "Episode: 991 \t\t Reward: -100.17299999999973\n",
      "Episode: 992 \t\t Reward: -100.07699999999976\n",
      "Episode: 993 \t\t Reward: -95.57199999999978\n",
      "Episode: 994 \t\t Reward: -101.60899999999978\n",
      "Episode: 995 \t\t Reward: -101.95399999999977\n",
      "Episode: 996 \t\t Reward: -102.06399999999977\n",
      "Episode: 997 \t\t Reward: -102.80349999999977\n",
      "Episode: 998 \t\t Reward: -102.77649999999976\n",
      "Episode: 999 \t\t Reward: -98.56849999999979\n",
      "Episode: 1000 \t\t Reward: -96.56299999999973\n",
      "Total test time: 02:24:57.51\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from collections import deque\n",
    "import numpy as np\n",
    "from utility import get_unique_numbers, count_edgezone\n",
    "import time\n",
    "# DDPG\n",
    "\n",
    "test_running_reward = 0 # reset calculation reward\n",
    "\n",
    "ac =[]\n",
    "start = time.time()\n",
    "runn_ep =[]\n",
    "scores =[]\n",
    "scores_deque = deque(maxlen=100)\n",
    "avg_scores =[]\n",
    "Bmp = []\n",
    "T_Bmp = []\n",
    "bumps = 0\n",
    "if p_model.is_dir():\n",
    "    nb_name = 'DDPG_Test_log.txt'\n",
    "with open(p_model.joinpath(nb_name), \"w\") as f:\n",
    "\n",
    "    for ep in range(1, total_test_episodes+1):\n",
    "        \n",
    "        start_ep = time.time()\n",
    "        ep_reward = 0\n",
    "        runn_ep.append(ep)\n",
    "        state = env.reset()\n",
    "        Ep_bmp = []\n",
    "        for t in range(1, 1500):\n",
    "            # bumps = 0\n",
    "            action = agent_DDPG.act(state)\n",
    "            ac.append(action)\n",
    "            state, reward, done, bumps = env.step(action) # reward return\n",
    "\n",
    "            Ep_bmp.append(bumps)\n",
    "            T_Bmp.append(bumps)\n",
    "            ep_reward += reward\n",
    "            # env.render('rgb_array') # uncommand to display -- warning: expensive resources.\n",
    "\n",
    "            if done:\n",
    "                break\n",
    "        Bmp.append(Ep_bmp)\n",
    "        scores_deque.append(ep_reward)\n",
    "        avg_scores.append(np.mean(scores_deque))\n",
    "        end_ep = time.time()\n",
    "        hours_ep, rem_ep = divmod(end_ep-start_ep, 3600)\n",
    "        minutes_ep, seconds_ep = divmod(rem_ep, 60)\n",
    "        \n",
    "\n",
    "        f.write(\"Episode{} (survive time): {:0>2}:{:0>2}:{:05.2f}\\n\".format(ep,int(hours_ep),int(minutes_ep),seconds_ep))\n",
    "        test_running_reward +=  ep_reward\n",
    "\n",
    "        # print('Episode: {} \\t\\t Reward: {}'.format(ep, round(ep_reward, 2)))\n",
    "        print('Episode: {} \\t\\t Reward: {}'.format(ep, avg_scores[ep-1]))\n",
    "\n",
    "        scores.append(round(ep_reward, 2))\n",
    "        ep_reward = 0\n",
    "    env.close()\n",
    "\n",
    "    end = time.time()\n",
    "    hours, rem = divmod(end-start, 3600)\n",
    "    minutes, seconds = divmod(rem, 60)\n",
    "    print(\"Total test time: {:0>2}:{:0>2}:{:05.2f}\".format(int(hours),int(minutes),seconds))\n",
    "    \n",
    "    # count the bump instances.\n",
    "    count_list_DDPG = []\n",
    "    for i in range(len(Bmp)):\n",
    "        count_list_DDPG.append(count_impulses(Bmp[i]))\n",
    "    data_ddpg = {'Bump_instance': count_list_DDPG}\n",
    "    Bump_track = pd.DataFrame(data_ddpg)\n",
    "    Tcount_DDPG = count_edgezone(T_Bmp,gap_threshold )\n",
    "    \n",
    "    f.write(\"Total Test time (survive time): {:0>2}:{:0>2}:{:05.2f}\\n\".format(int(hours),int(minutes),seconds))\n",
    "    f.write('Toal epoch:%s\\n'%total_test_episodes)\n",
    "    f.write(\"Bump instances of total: %s\\n\"%Tcount_DDPG)\n",
    "    f.write(\"Bump to edge zone %s\\n\"%count_edgezone(T_Bmp, gap_threshold))\n",
    "    f.close()\n",
    "\n",
    "\n",
    "DDPG_avg = avg_scores\n",
    "DDPG_scores = scores\n",
    "DDPG_ac = ac\n",
    "\n",
    "# save csv file\n",
    "csv_name = f\"DDPG_Reshape\"\n",
    "score_action_save(DDPG_ac,DDPG_scores,DDPG_avg,file_name=csv_name,path=p_model,print_=False)\n",
    "# score_save(scores,avg_scores,file_name=csv_name,path=p_model,print_=False) # save score table as csv file\n",
    "count_impulse_file = f\"DDPG_Reshape_count\"\n",
    "Bump_track.to_csv((p_model/f\"{count_impulse_file}.csv\")) # save bump instances as csv file\n",
    "\n",
    "\n",
    "# # plot the bump instances.\n",
    "# for i in range(total_test_episodes):\n",
    "#     plt.subplot(1, total_test_episodes, i+1)\n",
    "#     plt.plot(Bmp[i])\n",
    "#     plt.title(f'Ct_:{data[\"Bump_instance\"][i]}')\n",
    "\n",
    "# plot for touch instanaces for the track.\n",
    "# fig = plt.figure()\n",
    "# plt.plot(T_Bmp)\n",
    "# plt.title(f'Total bump instances:{Tcount_DDPG}')\n",
    "\n",
    "fig = plt.figure()\n",
    "ax = fig.add_subplot(111)\n",
    "plt.plot(avg_scores)\n",
    "# plt.plot(df['Averagescores'])\n",
    "plt.ylabel('Scores')\n",
    "plt.xlabel(f'Episode {len(avg_scores)}')\n",
    "plt.show()\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "\n",
    "action_list = [0,1,2,3,4]\n",
    "h=[]\n",
    "for i in action_list:\n",
    "    h.append(ac.count(i))\n",
    "bar_contain = plt.bar(action_list,h)\n",
    "plt.xlabel('Action items')\n",
    "plt.ylabel('Number of action item used during trial')\n",
    "plt.bar_label(bar_contain)\n",
    "plt.show()\n",
    "\n",
    "\n"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## DQN Reshape agent"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import gc\n",
    "\n",
    "# del agent_DQN\n",
    "del agent_DDPG\n",
    "# del agent_PPO\n",
    "torch.cuda.empty_cache()\n",
    "gc.collect()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Episode: 1 \t\t Reward: -27.850000000000122\n",
      "Episode: 2 \t\t Reward: -27.925000000000125\n",
      "Episode: 3 \t\t Reward: -27.900000000000123\n",
      "Episode: 4 \t\t Reward: -27.887500000000124\n",
      "Episode: 5 \t\t Reward: -27.910000000000128\n",
      "Episode: 6 \t\t Reward: -27.90000000000013\n",
      "Episode: 7 \t\t Reward: -27.914285714285842\n",
      "Episode: 8 \t\t Reward: -27.906250000000124\n",
      "Episode: 9 \t\t Reward: -27.900000000000123\n",
      "Episode: 10 \t\t Reward: -25.09500000000011\n",
      "Episode: 11 \t\t Reward: -25.359090909091023\n",
      "Episode: 12 \t\t Reward: -25.56666666666678\n",
      "Episode: 13 \t\t Reward: -25.75384615384627\n",
      "Episode: 14 \t\t Reward: -25.91428571428583\n",
      "Episode: 15 \t\t Reward: -26.04333333333345\n",
      "Episode: 16 \t\t Reward: -26.156250000000117\n",
      "Episode: 17 \t\t Reward: -26.26470588235306\n",
      "Episode: 18 \t\t Reward: -26.352777777777895\n",
      "Episode: 19 \t\t Reward: -26.43157894736854\n",
      "Episode: 20 \t\t Reward: -26.50250000000012\n",
      "Episode: 21 \t\t Reward: -25.233333333333444\n",
      "Episode: 22 \t\t Reward: -25.35227272727284\n",
      "Episode: 23 \t\t Reward: -25.460869565217507\n",
      "Episode: 24 \t\t Reward: -25.560416666666782\n",
      "Episode: 25 \t\t Reward: -25.652000000000115\n",
      "Episode: 26 \t\t Reward: -25.73653846153858\n",
      "Episode: 27 \t\t Reward: -25.820370370370487\n",
      "Episode: 28 \t\t Reward: -25.89285714285726\n",
      "Episode: 29 \t\t Reward: -25.960344827586326\n",
      "Episode: 30 \t\t Reward: -26.02333333333345\n",
      "Episode: 31 \t\t Reward: -26.08225806451625\n",
      "Episode: 32 \t\t Reward: -26.137500000000117\n",
      "Episode: 33 \t\t Reward: -26.189393939394055\n",
      "Episode: 34 \t\t Reward: -26.238235294117764\n",
      "Episode: 35 \t\t Reward: -26.284285714285833\n",
      "Episode: 36 \t\t Reward: -26.327777777777897\n",
      "Episode: 37 \t\t Reward: -26.36891891891904\n",
      "Episode: 38 \t\t Reward: -26.407894736842223\n",
      "Episode: 39 \t\t Reward: -26.444871794871915\n",
      "Episode: 40 \t\t Reward: -26.48000000000012\n",
      "Episode: 41 \t\t Reward: -26.51341463414646\n",
      "Episode: 42 \t\t Reward: -26.54880952380965\n",
      "Episode: 43 \t\t Reward: -26.579069767441982\n",
      "Episode: 44 \t\t Reward: -26.60795454545467\n",
      "Episode: 45 \t\t Reward: -26.63555555555568\n",
      "Episode: 46 \t\t Reward: -26.661956521739256\n",
      "Episode: 47 \t\t Reward: -26.687234042553317\n",
      "Episode: 48 \t\t Reward: -26.71458333333345\n",
      "Episode: 49 \t\t Reward: -26.740816326530734\n",
      "Episode: 50 \t\t Reward: -26.76300000000012\n",
      "Episode: 51 \t\t Reward: -26.235294117647175\n",
      "Episode: 52 \t\t Reward: -26.26634615384627\n",
      "Episode: 53 \t\t Reward: -25.767924528302\n",
      "Episode: 54 \t\t Reward: -25.809259259259374\n",
      "Episode: 55 \t\t Reward: -25.84636363636375\n",
      "Episode: 56 \t\t Reward: -25.882142857142973\n",
      "Episode: 57 \t\t Reward: -25.91666666666678\n",
      "Episode: 58 \t\t Reward: -25.950000000000117\n",
      "Episode: 59 \t\t Reward: -25.982203389830627\n",
      "Episode: 60 \t\t Reward: -26.01333333333345\n",
      "Episode: 61 \t\t Reward: -25.58442622950831\n",
      "Episode: 62 \t\t Reward: -25.620967741935598\n",
      "Episode: 63 \t\t Reward: -25.656349206349322\n",
      "Episode: 64 \t\t Reward: -25.690625000000118\n",
      "Episode: 65 \t\t Reward: -25.726153846153966\n",
      "Episode: 66 \t\t Reward: -25.758333333333454\n",
      "Episode: 67 \t\t Reward: -25.789552238806092\n",
      "Episode: 68 \t\t Reward: -25.819852941176592\n",
      "Episode: 69 \t\t Reward: -25.849275362318963\n",
      "Episode: 70 \t\t Reward: -25.877857142857263\n",
      "Episode: 71 \t\t Reward: -25.90563380281702\n",
      "Episode: 72 \t\t Reward: -25.932638888889002\n",
      "Episode: 73 \t\t Reward: -25.958904109589156\n",
      "Episode: 74 \t\t Reward: -25.984459459459575\n",
      "Episode: 75 \t\t Reward: -26.01133333333345\n",
      "Episode: 76 \t\t Reward: -26.035526315789593\n",
      "Episode: 77 \t\t Reward: -26.059090909091026\n",
      "Episode: 78 \t\t Reward: -26.083974358974476\n",
      "Episode: 79 \t\t Reward: -26.108227848101382\n",
      "Episode: 80 \t\t Reward: -26.13000000000012\n",
      "Episode: 81 \t\t Reward: -26.153086419753205\n",
      "Episode: 82 \t\t Reward: -26.173780487804994\n",
      "Episode: 83 \t\t Reward: -26.19397590361457\n",
      "Episode: 84 \t\t Reward: -26.21369047619059\n",
      "Episode: 85 \t\t Reward: -26.23470588235305\n",
      "Episode: 86 \t\t Reward: -26.25348837209313\n",
      "Episode: 87 \t\t Reward: -26.271839080459877\n",
      "Episode: 88 \t\t Reward: -26.28977272727285\n",
      "Episode: 89 \t\t Reward: -26.308988764045065\n",
      "Episode: 90 \t\t Reward: -26.32611111111123\n",
      "Episode: 91 \t\t Reward: -26.342857142857262\n",
      "Episode: 92 \t\t Reward: -26.359239130434897\n",
      "Episode: 93 \t\t Reward: -26.375268817204415\n",
      "Episode: 94 \t\t Reward: -26.39095744680862\n",
      "Episode: 95 \t\t Reward: -26.406315789473794\n",
      "Episode: 96 \t\t Reward: -26.421354166666788\n",
      "Episode: 97 \t\t Reward: -26.43608247422692\n",
      "Episode: 98 \t\t Reward: -26.452040816326647\n",
      "Episode: 99 \t\t Reward: -26.428282828282946\n",
      "Episode: 100 \t\t Reward: -26.442500000000113\n",
      "Episode: 101 \t\t Reward: -26.442500000000113\n",
      "Episode: 102 \t\t Reward: -26.441000000000113\n",
      "Episode: 103 \t\t Reward: -26.441000000000113\n",
      "Episode: 104 \t\t Reward: -26.441000000000113\n",
      "Episode: 105 \t\t Reward: -26.441000000000113\n",
      "Episode: 106 \t\t Reward: -26.16100000000011\n",
      "Episode: 107 \t\t Reward: -26.15950000000011\n",
      "Episode: 108 \t\t Reward: -26.15950000000011\n",
      "Episode: 109 \t\t Reward: -26.15950000000011\n",
      "Episode: 110 \t\t Reward: -26.439500000000113\n",
      "Episode: 111 \t\t Reward: -26.438000000000113\n",
      "Episode: 112 \t\t Reward: -26.438000000000113\n",
      "Episode: 113 \t\t Reward: -26.436500000000116\n",
      "Episode: 114 \t\t Reward: -26.43500000000011\n",
      "Episode: 115 \t\t Reward: -26.43500000000011\n",
      "Episode: 116 \t\t Reward: -26.43500000000011\n",
      "Episode: 117 \t\t Reward: -26.433500000000112\n",
      "Episode: 118 \t\t Reward: -26.435000000000112\n",
      "Episode: 119 \t\t Reward: -26.435000000000112\n",
      "Episode: 120 \t\t Reward: -26.43500000000011\n",
      "Episode: 121 \t\t Reward: -26.716500000000114\n",
      "Episode: 122 \t\t Reward: -26.71650000000011\n",
      "Episode: 123 \t\t Reward: -26.718000000000117\n",
      "Episode: 124 \t\t Reward: -26.71800000000011\n",
      "Episode: 125 \t\t Reward: -26.718000000000117\n",
      "Episode: 126 \t\t Reward: -26.718000000000117\n",
      "Episode: 127 \t\t Reward: -26.718000000000117\n",
      "Episode: 128 \t\t Reward: -26.71800000000011\n",
      "Episode: 129 \t\t Reward: -26.718000000000117\n",
      "Episode: 130 \t\t Reward: -26.718000000000117\n",
      "Episode: 131 \t\t Reward: -26.71800000000011\n",
      "Episode: 132 \t\t Reward: -26.71950000000011\n",
      "Episode: 133 \t\t Reward: -26.71950000000011\n",
      "Episode: 134 \t\t Reward: -26.439500000000113\n",
      "Episode: 135 \t\t Reward: -26.439500000000113\n",
      "Episode: 136 \t\t Reward: -26.439500000000113\n",
      "Episode: 137 \t\t Reward: -26.43950000000011\n",
      "Episode: 138 \t\t Reward: -26.441000000000113\n",
      "Episode: 139 \t\t Reward: -26.441000000000113\n",
      "Episode: 140 \t\t Reward: -26.441000000000113\n",
      "Episode: 141 \t\t Reward: -26.44250000000011\n",
      "Episode: 142 \t\t Reward: -26.441000000000113\n",
      "Episode: 143 \t\t Reward: -26.441000000000113\n",
      "Episode: 144 \t\t Reward: -26.441000000000113\n",
      "Episode: 145 \t\t Reward: -26.441000000000113\n",
      "Episode: 146 \t\t Reward: -26.441000000000113\n",
      "Episode: 147 \t\t Reward: -26.441000000000113\n",
      "Episode: 148 \t\t Reward: -26.439500000000113\n",
      "Episode: 149 \t\t Reward: -26.438000000000113\n",
      "Episode: 150 \t\t Reward: -26.15800000000011\n",
      "Episode: 151 \t\t Reward: -26.438000000000113\n",
      "Episode: 152 \t\t Reward: -26.438000000000113\n",
      "Episode: 153 \t\t Reward: -26.718000000000117\n",
      "Episode: 154 \t\t Reward: -26.716500000000114\n",
      "Episode: 155 \t\t Reward: -26.716500000000114\n",
      "Episode: 156 \t\t Reward: -26.71650000000011\n",
      "Episode: 157 \t\t Reward: -26.716500000000114\n",
      "Episode: 158 \t\t Reward: -26.716500000000114\n",
      "Episode: 159 \t\t Reward: -26.716500000000114\n",
      "Episode: 160 \t\t Reward: -26.716500000000114\n",
      "Episode: 161 \t\t Reward: -26.996500000000115\n",
      "Episode: 162 \t\t Reward: -26.996500000000115\n",
      "Episode: 163 \t\t Reward: -26.996500000000115\n",
      "Episode: 164 \t\t Reward: -26.996500000000115\n",
      "Episode: 165 \t\t Reward: -26.995000000000115\n",
      "Episode: 166 \t\t Reward: -26.995000000000115\n",
      "Episode: 167 \t\t Reward: -26.995000000000115\n",
      "Episode: 168 \t\t Reward: -26.995000000000115\n",
      "Episode: 169 \t\t Reward: -26.995000000000115\n",
      "Episode: 170 \t\t Reward: -26.995000000000115\n",
      "Episode: 171 \t\t Reward: -26.995000000000115\n",
      "Episode: 172 \t\t Reward: -26.995000000000115\n",
      "Episode: 173 \t\t Reward: -26.995000000000115\n",
      "Episode: 174 \t\t Reward: -26.995000000000115\n",
      "Episode: 175 \t\t Reward: -26.99350000000011\n",
      "Episode: 176 \t\t Reward: -26.99350000000011\n",
      "Episode: 177 \t\t Reward: -26.99350000000011\n",
      "Episode: 178 \t\t Reward: -26.99200000000011\n",
      "Episode: 179 \t\t Reward: -26.990500000000115\n",
      "Episode: 180 \t\t Reward: -26.990500000000115\n",
      "Episode: 181 \t\t Reward: -26.989000000000114\n",
      "Episode: 182 \t\t Reward: -26.989000000000114\n",
      "Episode: 183 \t\t Reward: -26.990500000000115\n",
      "Episode: 184 \t\t Reward: -26.992000000000118\n",
      "Episode: 185 \t\t Reward: -26.99050000000012\n",
      "Episode: 186 \t\t Reward: -26.99050000000012\n",
      "Episode: 187 \t\t Reward: -26.990500000000115\n",
      "Episode: 188 \t\t Reward: -26.990500000000115\n",
      "Episode: 189 \t\t Reward: -26.990500000000115\n",
      "Episode: 190 \t\t Reward: -26.992000000000118\n",
      "Episode: 191 \t\t Reward: -26.992000000000118\n",
      "Episode: 192 \t\t Reward: -26.992000000000118\n",
      "Episode: 193 \t\t Reward: -26.99200000000011\n",
      "Episode: 194 \t\t Reward: -26.992000000000118\n",
      "Episode: 195 \t\t Reward: -26.993500000000118\n",
      "Episode: 196 \t\t Reward: -26.99500000000012\n",
      "Episode: 197 \t\t Reward: -26.99650000000012\n",
      "Episode: 198 \t\t Reward: -26.99650000000012\n",
      "Episode: 199 \t\t Reward: -27.03400000000012\n",
      "Episode: 200 \t\t Reward: -27.03400000000012\n",
      "Episode: 201 \t\t Reward: -27.034000000000116\n",
      "Episode: 202 \t\t Reward: -27.034000000000116\n",
      "Episode: 203 \t\t Reward: -27.034000000000116\n",
      "Episode: 204 \t\t Reward: -27.034000000000116\n",
      "Episode: 205 \t\t Reward: -27.032500000000113\n",
      "Episode: 206 \t\t Reward: -27.312500000000114\n",
      "Episode: 207 \t\t Reward: -27.312500000000117\n",
      "Episode: 208 \t\t Reward: -27.312500000000117\n",
      "Episode: 209 \t\t Reward: -27.31400000000012\n",
      "Episode: 210 \t\t Reward: -27.314000000000114\n",
      "Episode: 211 \t\t Reward: -27.03400000000012\n",
      "Episode: 212 \t\t Reward: -27.03550000000012\n",
      "Episode: 213 \t\t Reward: -27.035500000000116\n",
      "Episode: 214 \t\t Reward: -27.035500000000116\n",
      "Episode: 215 \t\t Reward: -27.035500000000116\n",
      "Episode: 216 \t\t Reward: -27.035500000000113\n",
      "Episode: 217 \t\t Reward: -27.035500000000116\n",
      "Episode: 218 \t\t Reward: -27.034000000000116\n",
      "Episode: 219 \t\t Reward: -27.034000000000116\n",
      "Episode: 220 \t\t Reward: -26.754000000000115\n",
      "Episode: 221 \t\t Reward: -26.754000000000115\n",
      "Episode: 222 \t\t Reward: -26.754000000000115\n",
      "Episode: 223 \t\t Reward: -26.752500000000115\n",
      "Episode: 224 \t\t Reward: -26.752500000000115\n",
      "Episode: 225 \t\t Reward: -26.752500000000115\n",
      "Episode: 226 \t\t Reward: -26.754000000000115\n",
      "Episode: 227 \t\t Reward: -26.752500000000115\n",
      "Episode: 228 \t\t Reward: -26.752500000000115\n",
      "Episode: 229 \t\t Reward: -26.75400000000012\n",
      "Episode: 230 \t\t Reward: -26.754000000000115\n",
      "Episode: 231 \t\t Reward: -26.754000000000115\n",
      "Episode: 232 \t\t Reward: -26.752500000000115\n",
      "Episode: 233 \t\t Reward: -26.754000000000115\n",
      "Episode: 234 \t\t Reward: -27.034000000000116\n",
      "Episode: 235 \t\t Reward: -27.03400000000012\n",
      "Episode: 236 \t\t Reward: -27.03400000000012\n",
      "Episode: 237 \t\t Reward: -27.034000000000116\n",
      "Episode: 238 \t\t Reward: -27.032500000000113\n",
      "Episode: 239 \t\t Reward: -27.03250000000012\n",
      "Episode: 240 \t\t Reward: -27.032500000000113\n",
      "Episode: 241 \t\t Reward: -27.031000000000116\n",
      "Episode: 242 \t\t Reward: -27.031000000000112\n",
      "Episode: 243 \t\t Reward: -27.031000000000116\n",
      "Episode: 244 \t\t Reward: -27.031000000000112\n",
      "Episode: 245 \t\t Reward: -27.031000000000116\n",
      "Episode: 246 \t\t Reward: -27.032500000000113\n",
      "Episode: 247 \t\t Reward: -27.03400000000012\n",
      "Episode: 248 \t\t Reward: -27.035500000000116\n",
      "Episode: 249 \t\t Reward: -27.03550000000012\n",
      "Episode: 250 \t\t Reward: -27.315500000000114\n",
      "Episode: 251 \t\t Reward: -27.315500000000114\n",
      "Episode: 252 \t\t Reward: -27.315500000000114\n",
      "Episode: 253 \t\t Reward: -27.315500000000114\n",
      "Episode: 254 \t\t Reward: -27.315500000000114\n",
      "Episode: 255 \t\t Reward: -27.275000000000123\n",
      "Episode: 256 \t\t Reward: -27.27650000000012\n",
      "Episode: 257 \t\t Reward: -27.27800000000012\n",
      "Episode: 258 \t\t Reward: -27.278000000000123\n",
      "Episode: 259 \t\t Reward: -27.27950000000012\n",
      "Episode: 260 \t\t Reward: -27.279500000000116\n",
      "Episode: 261 \t\t Reward: -27.281000000000116\n",
      "Episode: 262 \t\t Reward: -27.281000000000112\n",
      "Episode: 263 \t\t Reward: -27.28250000000012\n",
      "Episode: 264 \t\t Reward: -27.28250000000012\n",
      "Episode: 265 \t\t Reward: -27.282500000000113\n",
      "Episode: 266 \t\t Reward: -27.282500000000113\n",
      "Episode: 267 \t\t Reward: -27.282500000000113\n",
      "Episode: 268 \t\t Reward: -27.282500000000113\n",
      "Episode: 269 \t\t Reward: -27.282500000000113\n",
      "Episode: 270 \t\t Reward: -27.282500000000113\n",
      "Episode: 271 \t\t Reward: -27.282500000000113\n",
      "Episode: 272 \t\t Reward: -27.282500000000113\n",
      "Episode: 273 \t\t Reward: -27.28250000000012\n",
      "Episode: 274 \t\t Reward: -27.282500000000113\n",
      "Episode: 275 \t\t Reward: -27.249500000000122\n",
      "Episode: 276 \t\t Reward: -27.249500000000115\n",
      "Episode: 277 \t\t Reward: -26.96950000000012\n",
      "Episode: 278 \t\t Reward: -26.969500000000117\n",
      "Episode: 279 \t\t Reward: -26.969500000000117\n",
      "Episode: 280 \t\t Reward: -26.96950000000011\n",
      "Episode: 281 \t\t Reward: -26.969500000000117\n",
      "Episode: 282 \t\t Reward: -26.96950000000011\n",
      "Episode: 283 \t\t Reward: -26.96800000000011\n",
      "Episode: 284 \t\t Reward: -26.966500000000114\n",
      "Episode: 285 \t\t Reward: -26.966500000000114\n",
      "Episode: 286 \t\t Reward: -26.686500000000116\n",
      "Episode: 287 \t\t Reward: -26.686500000000116\n",
      "Episode: 288 \t\t Reward: -26.686500000000116\n",
      "Episode: 289 \t\t Reward: -26.405000000000115\n",
      "Episode: 290 \t\t Reward: -26.40350000000011\n",
      "Episode: 291 \t\t Reward: -26.40350000000011\n",
      "Episode: 292 \t\t Reward: -26.40350000000011\n",
      "Episode: 293 \t\t Reward: -26.40350000000011\n",
      "Episode: 294 \t\t Reward: -26.40350000000011\n",
      "Episode: 295 \t\t Reward: -26.40200000000011\n",
      "Episode: 296 \t\t Reward: -26.40050000000011\n",
      "Episode: 297 \t\t Reward: -26.40050000000011\n",
      "Episode: 298 \t\t Reward: -26.39900000000011\n",
      "Episode: 299 \t\t Reward: -26.39900000000011\n",
      "Episode: 300 \t\t Reward: -26.39900000000011\n",
      "Episode: 301 \t\t Reward: -26.39900000000011\n",
      "Episode: 302 \t\t Reward: -26.39900000000011\n",
      "Episode: 303 \t\t Reward: -26.39900000000011\n",
      "Episode: 304 \t\t Reward: -26.399000000000115\n",
      "Episode: 305 \t\t Reward: -26.39900000000011\n",
      "Episode: 306 \t\t Reward: -26.39900000000011\n",
      "Episode: 307 \t\t Reward: -26.39900000000011\n",
      "Episode: 308 \t\t Reward: -26.399000000000115\n",
      "Episode: 309 \t\t Reward: -26.39900000000011\n",
      "Episode: 310 \t\t Reward: -26.11900000000011\n",
      "Episode: 311 \t\t Reward: -26.399000000000115\n",
      "Episode: 312 \t\t Reward: -26.397500000000115\n",
      "Episode: 313 \t\t Reward: -26.397500000000107\n",
      "Episode: 314 \t\t Reward: -26.397500000000107\n",
      "Episode: 315 \t\t Reward: -26.397500000000107\n",
      "Episode: 316 \t\t Reward: -26.397500000000107\n",
      "Episode: 317 \t\t Reward: -26.397500000000107\n",
      "Episode: 318 \t\t Reward: -26.397500000000107\n",
      "Episode: 319 \t\t Reward: -26.39900000000011\n",
      "Episode: 320 \t\t Reward: -26.67900000000011\n",
      "Episode: 321 \t\t Reward: -26.679000000000116\n",
      "Episode: 322 \t\t Reward: -26.679000000000116\n",
      "Episode: 323 \t\t Reward: -26.679000000000116\n",
      "Episode: 324 \t\t Reward: -26.66900000000011\n",
      "Episode: 325 \t\t Reward: -26.66900000000011\n",
      "Episode: 326 \t\t Reward: -26.66750000000011\n",
      "Episode: 327 \t\t Reward: -26.667500000000114\n",
      "Episode: 328 \t\t Reward: -26.66750000000011\n",
      "Episode: 329 \t\t Reward: -26.666000000000107\n",
      "Episode: 330 \t\t Reward: -26.666000000000114\n",
      "Episode: 331 \t\t Reward: -26.667500000000114\n",
      "Episode: 332 \t\t Reward: -26.667500000000114\n",
      "Episode: 333 \t\t Reward: -26.666000000000107\n",
      "Episode: 334 \t\t Reward: -26.666000000000114\n",
      "Episode: 335 \t\t Reward: -26.666000000000114\n",
      "Episode: 336 \t\t Reward: -26.667500000000114\n",
      "Episode: 337 \t\t Reward: -26.667500000000114\n",
      "Episode: 338 \t\t Reward: -26.667500000000114\n",
      "Episode: 339 \t\t Reward: -26.667500000000114\n",
      "Episode: 340 \t\t Reward: -26.66750000000011\n",
      "Episode: 341 \t\t Reward: -26.66750000000011\n",
      "Episode: 342 \t\t Reward: -26.66750000000011\n",
      "Episode: 343 \t\t Reward: -26.66750000000011\n",
      "Episode: 344 \t\t Reward: -26.66750000000011\n",
      "Episode: 345 \t\t Reward: -26.38750000000011\n",
      "Episode: 346 \t\t Reward: -26.386000000000113\n",
      "Episode: 347 \t\t Reward: -26.384500000000113\n",
      "Episode: 348 \t\t Reward: -26.38300000000011\n",
      "Episode: 349 \t\t Reward: -26.38300000000011\n",
      "Episode: 350 \t\t Reward: -26.38300000000011\n",
      "Episode: 351 \t\t Reward: -26.38300000000011\n",
      "Episode: 352 \t\t Reward: -26.38300000000011\n",
      "Episode: 353 \t\t Reward: -26.38300000000011\n",
      "Episode: 354 \t\t Reward: -26.38300000000011\n",
      "Episode: 355 \t\t Reward: -26.423500000000114\n",
      "Episode: 356 \t\t Reward: -26.423500000000114\n",
      "Episode: 357 \t\t Reward: -26.42200000000011\n",
      "Episode: 358 \t\t Reward: -26.42200000000011\n",
      "Episode: 359 \t\t Reward: -26.422000000000118\n",
      "Episode: 360 \t\t Reward: -26.42200000000011\n",
      "Episode: 361 \t\t Reward: -26.420500000000114\n",
      "Episode: 362 \t\t Reward: -26.422000000000118\n",
      "Episode: 363 \t\t Reward: -26.420500000000114\n",
      "Episode: 364 \t\t Reward: -26.420500000000114\n",
      "Episode: 365 \t\t Reward: -26.420500000000114\n",
      "Episode: 366 \t\t Reward: -26.42050000000011\n",
      "Episode: 367 \t\t Reward: -26.42050000000011\n",
      "Episode: 368 \t\t Reward: -26.42050000000011\n",
      "Episode: 369 \t\t Reward: -26.42050000000011\n",
      "Episode: 370 \t\t Reward: -26.42050000000011\n",
      "Episode: 371 \t\t Reward: -26.42050000000011\n",
      "Episode: 372 \t\t Reward: -26.42050000000011\n",
      "Episode: 373 \t\t Reward: -26.42050000000011\n",
      "Episode: 374 \t\t Reward: -26.42050000000011\n",
      "Episode: 375 \t\t Reward: -26.455000000000112\n",
      "Episode: 376 \t\t Reward: -26.455000000000112\n",
      "Episode: 377 \t\t Reward: -26.735000000000113\n",
      "Episode: 378 \t\t Reward: -26.735000000000113\n",
      "Episode: 379 \t\t Reward: -26.735000000000117\n",
      "Episode: 380 \t\t Reward: -26.735000000000117\n",
      "Episode: 381 \t\t Reward: -26.736500000000113\n",
      "Episode: 382 \t\t Reward: -26.736500000000113\n",
      "Episode: 383 \t\t Reward: -26.73650000000012\n",
      "Episode: 384 \t\t Reward: -26.73650000000012\n",
      "Episode: 385 \t\t Reward: -26.736500000000113\n",
      "Episode: 386 \t\t Reward: -27.016500000000114\n",
      "Episode: 387 \t\t Reward: -27.018000000000114\n",
      "Episode: 388 \t\t Reward: -26.97750000000012\n",
      "Episode: 389 \t\t Reward: -27.257500000000118\n",
      "Episode: 390 \t\t Reward: -27.257500000000118\n",
      "Episode: 391 \t\t Reward: -27.25750000000012\n",
      "Episode: 392 \t\t Reward: -27.259000000000118\n",
      "Episode: 393 \t\t Reward: -27.259000000000114\n",
      "Episode: 394 \t\t Reward: -27.259000000000118\n",
      "Episode: 395 \t\t Reward: -27.259000000000118\n",
      "Episode: 396 \t\t Reward: -27.259000000000114\n",
      "Episode: 397 \t\t Reward: -27.259000000000114\n",
      "Episode: 398 \t\t Reward: -27.259000000000118\n",
      "Episode: 399 \t\t Reward: -27.260500000000114\n",
      "Episode: 400 \t\t Reward: -27.26200000000012\n",
      "Episode: 401 \t\t Reward: -27.263500000000118\n",
      "Episode: 402 \t\t Reward: -27.263500000000118\n",
      "Episode: 403 \t\t Reward: -27.263500000000114\n",
      "Episode: 404 \t\t Reward: -27.263500000000118\n",
      "Episode: 405 \t\t Reward: -27.263500000000114\n",
      "Episode: 406 \t\t Reward: -27.263500000000114\n",
      "Episode: 407 \t\t Reward: -27.263500000000114\n",
      "Episode: 408 \t\t Reward: -27.265000000000114\n",
      "Episode: 409 \t\t Reward: -27.263500000000118\n",
      "Episode: 410 \t\t Reward: -27.54350000000012\n",
      "Episode: 411 \t\t Reward: -27.54350000000012\n",
      "Episode: 412 \t\t Reward: -27.54350000000012\n",
      "Episode: 413 \t\t Reward: -27.54350000000012\n",
      "Episode: 414 \t\t Reward: -27.543500000000112\n",
      "Episode: 415 \t\t Reward: -27.54350000000012\n",
      "Episode: 416 \t\t Reward: -27.54500000000012\n",
      "Episode: 417 \t\t Reward: -27.54650000000012\n",
      "Episode: 418 \t\t Reward: -27.546500000000115\n",
      "Episode: 419 \t\t Reward: -27.54500000000012\n",
      "Episode: 420 \t\t Reward: -27.265000000000118\n",
      "Episode: 421 \t\t Reward: -27.263500000000114\n",
      "Episode: 422 \t\t Reward: -27.265000000000114\n",
      "Episode: 423 \t\t Reward: -27.265000000000114\n",
      "Episode: 424 \t\t Reward: -27.275000000000112\n",
      "Episode: 425 \t\t Reward: -27.275000000000112\n",
      "Episode: 426 \t\t Reward: -27.276500000000116\n",
      "Episode: 427 \t\t Reward: -27.276500000000116\n",
      "Episode: 428 \t\t Reward: -27.276500000000116\n",
      "Episode: 429 \t\t Reward: -27.276500000000116\n",
      "Episode: 430 \t\t Reward: -27.276500000000116\n",
      "Episode: 431 \t\t Reward: -27.275000000000112\n",
      "Episode: 432 \t\t Reward: -27.27500000000012\n",
      "Episode: 433 \t\t Reward: -27.275000000000112\n",
      "Episode: 434 \t\t Reward: -27.275000000000112\n",
      "Episode: 435 \t\t Reward: -27.27500000000012\n",
      "Episode: 436 \t\t Reward: -27.273500000000112\n",
      "Episode: 437 \t\t Reward: -27.273500000000116\n",
      "Episode: 438 \t\t Reward: -27.273500000000112\n",
      "Episode: 439 \t\t Reward: -27.273500000000116\n",
      "Episode: 440 \t\t Reward: -27.273500000000112\n",
      "Episode: 441 \t\t Reward: -27.273500000000112\n",
      "Episode: 442 \t\t Reward: -27.273500000000112\n",
      "Episode: 443 \t\t Reward: -27.273500000000112\n",
      "Episode: 444 \t\t Reward: -27.273500000000112\n",
      "Episode: 445 \t\t Reward: -27.553500000000113\n",
      "Episode: 446 \t\t Reward: -27.553500000000117\n",
      "Episode: 447 \t\t Reward: -27.273500000000112\n",
      "Episode: 448 \t\t Reward: -27.273500000000116\n",
      "Episode: 449 \t\t Reward: -27.273500000000116\n",
      "Episode: 450 \t\t Reward: -26.993500000000118\n",
      "Episode: 451 \t\t Reward: -26.99350000000011\n",
      "Episode: 452 \t\t Reward: -26.99500000000012\n",
      "Episode: 453 \t\t Reward: -26.99500000000012\n",
      "Episode: 454 \t\t Reward: -26.995000000000115\n",
      "Episode: 455 \t\t Reward: -26.995000000000115\n",
      "Episode: 456 \t\t Reward: -26.995000000000115\n",
      "Episode: 457 \t\t Reward: -26.995000000000115\n",
      "Episode: 458 \t\t Reward: -26.996500000000115\n",
      "Episode: 459 \t\t Reward: -26.99500000000012\n",
      "Episode: 460 \t\t Reward: -26.995000000000115\n",
      "Episode: 461 \t\t Reward: -26.995000000000115\n",
      "Episode: 462 \t\t Reward: -26.962000000000113\n",
      "Episode: 463 \t\t Reward: -26.962000000000117\n",
      "Episode: 464 \t\t Reward: -26.962000000000113\n",
      "Episode: 465 \t\t Reward: -26.962000000000113\n",
      "Episode: 466 \t\t Reward: -26.962000000000117\n",
      "Episode: 467 \t\t Reward: -26.962000000000113\n",
      "Episode: 468 \t\t Reward: -26.962000000000117\n",
      "Episode: 469 \t\t Reward: -26.963500000000113\n",
      "Episode: 470 \t\t Reward: -26.963500000000117\n",
      "Episode: 471 \t\t Reward: -26.963500000000113\n",
      "Episode: 472 \t\t Reward: -26.963500000000117\n",
      "Episode: 473 \t\t Reward: -26.963500000000113\n",
      "Episode: 474 \t\t Reward: -26.963500000000113\n",
      "Episode: 475 \t\t Reward: -26.962000000000113\n",
      "Episode: 476 \t\t Reward: -26.962000000000117\n",
      "Episode: 477 \t\t Reward: -26.960500000000117\n",
      "Episode: 478 \t\t Reward: -26.96200000000012\n",
      "Episode: 479 \t\t Reward: -26.962000000000117\n",
      "Episode: 480 \t\t Reward: -26.96350000000012\n",
      "Episode: 481 \t\t Reward: -26.962000000000117\n",
      "Episode: 482 \t\t Reward: -26.963500000000113\n",
      "Episode: 483 \t\t Reward: -26.963500000000113\n",
      "Episode: 484 \t\t Reward: -26.963500000000117\n",
      "Episode: 485 \t\t Reward: -26.963500000000117\n",
      "Episode: 486 \t\t Reward: -26.963500000000113\n",
      "Episode: 487 \t\t Reward: -26.962000000000117\n",
      "Episode: 488 \t\t Reward: -27.00250000000012\n",
      "Episode: 489 \t\t Reward: -27.002500000000115\n",
      "Episode: 490 \t\t Reward: -27.00250000000012\n",
      "Episode: 491 \t\t Reward: -27.004000000000115\n",
      "Episode: 492 \t\t Reward: -27.004000000000115\n",
      "Episode: 493 \t\t Reward: -27.004000000000115\n",
      "Episode: 494 \t\t Reward: -27.005500000000115\n",
      "Episode: 495 \t\t Reward: -27.005500000000115\n",
      "Episode: 496 \t\t Reward: -27.007000000000115\n",
      "Episode: 497 \t\t Reward: -27.00550000000012\n",
      "Episode: 498 \t\t Reward: -27.005500000000115\n",
      "Episode: 499 \t\t Reward: -27.00400000000012\n",
      "Episode: 500 \t\t Reward: -26.722500000000114\n",
      "Episode: 501 \t\t Reward: -26.721000000000114\n",
      "Episode: 502 \t\t Reward: -26.721000000000114\n",
      "Episode: 503 \t\t Reward: -26.721000000000114\n",
      "Episode: 504 \t\t Reward: -26.722500000000114\n",
      "Episode: 505 \t\t Reward: -26.724000000000114\n",
      "Episode: 506 \t\t Reward: -26.725500000000114\n",
      "Episode: 507 \t\t Reward: -26.725500000000114\n",
      "Episode: 508 \t\t Reward: -26.724000000000114\n",
      "Episode: 509 \t\t Reward: -26.72400000000011\n",
      "Episode: 510 \t\t Reward: -26.72400000000011\n",
      "Episode: 511 \t\t Reward: -26.72400000000011\n",
      "Episode: 512 \t\t Reward: -26.725500000000114\n",
      "Episode: 513 \t\t Reward: -26.725500000000114\n",
      "Episode: 514 \t\t Reward: -26.72700000000011\n",
      "Episode: 515 \t\t Reward: -26.72700000000011\n",
      "Episode: 516 \t\t Reward: -26.727000000000118\n",
      "Episode: 517 \t\t Reward: -26.725500000000114\n",
      "Episode: 518 \t\t Reward: -26.72550000000011\n",
      "Episode: 519 \t\t Reward: -26.72550000000011\n",
      "Episode: 520 \t\t Reward: -27.00700000000011\n",
      "Episode: 521 \t\t Reward: -27.00700000000011\n",
      "Episode: 522 \t\t Reward: -27.005500000000115\n",
      "Episode: 523 \t\t Reward: -27.00550000000011\n",
      "Episode: 524 \t\t Reward: -27.00550000000011\n",
      "Episode: 525 \t\t Reward: -27.00550000000011\n",
      "Episode: 526 \t\t Reward: -27.004000000000115\n",
      "Episode: 527 \t\t Reward: -27.00550000000011\n",
      "Episode: 528 \t\t Reward: -27.005500000000115\n",
      "Episode: 529 \t\t Reward: -26.725500000000114\n",
      "Episode: 530 \t\t Reward: -26.727000000000118\n",
      "Episode: 531 \t\t Reward: -26.72700000000011\n",
      "Episode: 532 \t\t Reward: -26.72700000000011\n",
      "Episode: 533 \t\t Reward: -26.72700000000011\n",
      "Episode: 534 \t\t Reward: -26.727000000000107\n",
      "Episode: 535 \t\t Reward: -26.72700000000011\n",
      "Episode: 536 \t\t Reward: -26.727000000000107\n",
      "Episode: 537 \t\t Reward: -26.727000000000107\n",
      "Episode: 538 \t\t Reward: -26.72700000000011\n",
      "Episode: 539 \t\t Reward: -26.72700000000011\n",
      "Episode: 540 \t\t Reward: -26.72700000000011\n",
      "Episode: 541 \t\t Reward: -26.72700000000011\n",
      "Episode: 542 \t\t Reward: -26.447000000000116\n",
      "Episode: 543 \t\t Reward: -26.447000000000113\n",
      "Episode: 544 \t\t Reward: -26.448500000000113\n",
      "Episode: 545 \t\t Reward: -26.448500000000113\n",
      "Episode: 546 \t\t Reward: -26.448500000000113\n",
      "Episode: 547 \t\t Reward: -26.728500000000114\n",
      "Episode: 548 \t\t Reward: -26.728500000000114\n",
      "Episode: 549 \t\t Reward: -26.728500000000114\n",
      "Episode: 550 \t\t Reward: -27.00850000000011\n",
      "Episode: 551 \t\t Reward: -27.00850000000011\n",
      "Episode: 552 \t\t Reward: -27.007000000000115\n",
      "Episode: 553 \t\t Reward: -27.007000000000115\n",
      "Episode: 554 \t\t Reward: -27.007000000000115\n",
      "Episode: 555 \t\t Reward: -27.007000000000115\n",
      "Episode: 556 \t\t Reward: -27.00700000000011\n",
      "Episode: 557 \t\t Reward: -27.007000000000115\n",
      "Episode: 558 \t\t Reward: -27.005500000000115\n",
      "Episode: 559 \t\t Reward: -27.005500000000115\n",
      "Episode: 560 \t\t Reward: -27.00700000000011\n",
      "Episode: 561 \t\t Reward: -27.00850000000011\n",
      "Episode: 562 \t\t Reward: -27.04000000000012\n",
      "Episode: 563 \t\t Reward: -27.04000000000012\n",
      "Episode: 564 \t\t Reward: -27.040000000000113\n",
      "Episode: 565 \t\t Reward: -27.040000000000113\n",
      "Episode: 566 \t\t Reward: -27.04000000000012\n",
      "Episode: 567 \t\t Reward: -27.04000000000012\n",
      "Episode: 568 \t\t Reward: -27.04150000000012\n",
      "Episode: 569 \t\t Reward: -27.040000000000113\n",
      "Episode: 570 \t\t Reward: -27.040000000000113\n",
      "Episode: 571 \t\t Reward: -27.04000000000012\n",
      "Episode: 572 \t\t Reward: -27.04000000000012\n",
      "Episode: 573 \t\t Reward: -27.04000000000012\n",
      "Episode: 574 \t\t Reward: -27.04000000000012\n",
      "Episode: 575 \t\t Reward: -27.04000000000012\n",
      "Episode: 576 \t\t Reward: -27.04000000000012\n",
      "Episode: 577 \t\t Reward: -27.00850000000012\n",
      "Episode: 578 \t\t Reward: -27.007000000000122\n",
      "Episode: 579 \t\t Reward: -27.007000000000122\n",
      "Episode: 580 \t\t Reward: -27.005500000000115\n",
      "Episode: 581 \t\t Reward: -27.00550000000011\n",
      "Episode: 582 \t\t Reward: -27.005500000000115\n",
      "Episode: 583 \t\t Reward: -26.969500000000117\n",
      "Episode: 584 \t\t Reward: -26.969500000000117\n",
      "Episode: 585 \t\t Reward: -26.971000000000117\n",
      "Episode: 586 \t\t Reward: -26.971000000000117\n",
      "Episode: 587 \t\t Reward: -26.971000000000117\n",
      "Episode: 588 \t\t Reward: -26.972500000000117\n",
      "Episode: 589 \t\t Reward: -26.972500000000114\n",
      "Episode: 590 \t\t Reward: -26.974000000000114\n",
      "Episode: 591 \t\t Reward: -26.972500000000117\n",
      "Episode: 592 \t\t Reward: -26.971000000000117\n",
      "Episode: 593 \t\t Reward: -26.971000000000117\n",
      "Episode: 594 \t\t Reward: -26.969500000000117\n",
      "Episode: 595 \t\t Reward: -26.969500000000117\n",
      "Episode: 596 \t\t Reward: -26.96800000000011\n",
      "Episode: 597 \t\t Reward: -26.96950000000011\n",
      "Episode: 598 \t\t Reward: -26.969500000000117\n",
      "Episode: 599 \t\t Reward: -26.96950000000011\n",
      "Episode: 600 \t\t Reward: -27.25100000000011\n",
      "Episode: 601 \t\t Reward: -27.25250000000012\n",
      "Episode: 602 \t\t Reward: -27.25250000000012\n",
      "Episode: 603 \t\t Reward: -27.25250000000012\n",
      "Episode: 604 \t\t Reward: -27.25100000000011\n",
      "Episode: 605 \t\t Reward: -27.25100000000012\n",
      "Episode: 606 \t\t Reward: -27.24950000000011\n",
      "Episode: 607 \t\t Reward: -27.249500000000115\n",
      "Episode: 608 \t\t Reward: -27.24950000000011\n",
      "Episode: 609 \t\t Reward: -27.249500000000115\n",
      "Episode: 610 \t\t Reward: -26.969500000000117\n",
      "Episode: 611 \t\t Reward: -26.969500000000117\n",
      "Episode: 612 \t\t Reward: -26.968000000000117\n",
      "Episode: 613 \t\t Reward: -26.968000000000117\n",
      "Episode: 614 \t\t Reward: -26.966500000000114\n",
      "Episode: 615 \t\t Reward: -26.966500000000114\n",
      "Episode: 616 \t\t Reward: -26.96650000000012\n",
      "Episode: 617 \t\t Reward: -26.966500000000114\n",
      "Episode: 618 \t\t Reward: -26.96650000000012\n",
      "Episode: 619 \t\t Reward: -26.966500000000114\n",
      "Episode: 620 \t\t Reward: -26.96650000000012\n",
      "Episode: 621 \t\t Reward: -26.966500000000114\n",
      "Episode: 622 \t\t Reward: -26.96650000000012\n",
      "Episode: 623 \t\t Reward: -26.968000000000117\n",
      "Episode: 624 \t\t Reward: -26.968000000000117\n",
      "Episode: 625 \t\t Reward: -26.969500000000117\n",
      "Episode: 626 \t\t Reward: -26.969500000000117\n",
      "Episode: 627 \t\t Reward: -26.969500000000117\n",
      "Episode: 628 \t\t Reward: -26.969500000000117\n",
      "Episode: 629 \t\t Reward: -27.25100000000012\n",
      "Episode: 630 \t\t Reward: -27.249500000000115\n",
      "Episode: 631 \t\t Reward: -27.249500000000115\n",
      "Episode: 632 \t\t Reward: -27.25100000000011\n",
      "Episode: 633 \t\t Reward: -27.25100000000011\n",
      "Episode: 634 \t\t Reward: -27.25100000000012\n",
      "Episode: 635 \t\t Reward: -27.25100000000012\n",
      "Episode: 636 \t\t Reward: -27.252500000000115\n",
      "Episode: 637 \t\t Reward: -27.252500000000115\n",
      "Episode: 638 \t\t Reward: -27.252500000000115\n",
      "Episode: 639 \t\t Reward: -27.252500000000115\n",
      "Episode: 640 \t\t Reward: -27.252500000000115\n",
      "Episode: 641 \t\t Reward: -27.252500000000115\n",
      "Episode: 642 \t\t Reward: -27.534000000000116\n",
      "Episode: 643 \t\t Reward: -27.534000000000116\n",
      "Episode: 644 \t\t Reward: -27.534000000000116\n",
      "Episode: 645 \t\t Reward: -27.534000000000116\n",
      "Episode: 646 \t\t Reward: -27.53400000000012\n",
      "Episode: 647 \t\t Reward: -27.50250000000012\n",
      "Episode: 648 \t\t Reward: -27.502500000000115\n",
      "Episode: 649 \t\t Reward: -27.504000000000115\n",
      "Episode: 650 \t\t Reward: -27.504000000000115\n",
      "Episode: 651 \t\t Reward: -27.50400000000012\n",
      "Episode: 652 \t\t Reward: -27.505500000000115\n",
      "Episode: 653 \t\t Reward: -27.505500000000115\n",
      "Episode: 654 \t\t Reward: -27.507000000000115\n",
      "Episode: 655 \t\t Reward: -27.507000000000115\n",
      "Episode: 656 \t\t Reward: -27.50550000000011\n",
      "Episode: 657 \t\t Reward: -27.505500000000115\n",
      "Episode: 658 \t\t Reward: -27.505500000000115\n",
      "Episode: 659 \t\t Reward: -27.50550000000011\n",
      "Episode: 660 \t\t Reward: -27.505500000000115\n",
      "Episode: 661 \t\t Reward: -27.504000000000115\n",
      "Episode: 662 \t\t Reward: -27.50400000000012\n",
      "Episode: 663 \t\t Reward: -27.504000000000115\n",
      "Episode: 664 \t\t Reward: -27.504000000000115\n",
      "Episode: 665 \t\t Reward: -27.504000000000115\n",
      "Episode: 666 \t\t Reward: -27.504000000000115\n",
      "Episode: 667 \t\t Reward: -27.504000000000115\n",
      "Episode: 668 \t\t Reward: -27.504000000000115\n",
      "Episode: 669 \t\t Reward: -27.504000000000115\n",
      "Episode: 670 \t\t Reward: -27.504000000000115\n",
      "Episode: 671 \t\t Reward: -27.505500000000115\n",
      "Episode: 672 \t\t Reward: -27.505500000000115\n",
      "Episode: 673 \t\t Reward: -27.505500000000115\n",
      "Episode: 674 \t\t Reward: -27.505500000000115\n",
      "Episode: 675 \t\t Reward: -27.50550000000011\n",
      "Episode: 676 \t\t Reward: -27.22550000000011\n",
      "Episode: 677 \t\t Reward: -27.25850000000011\n",
      "Episode: 678 \t\t Reward: -26.978500000000114\n",
      "Episode: 679 \t\t Reward: -26.978500000000114\n",
      "Episode: 680 \t\t Reward: -26.978500000000114\n",
      "Episode: 681 \t\t Reward: -26.978500000000114\n",
      "Episode: 682 \t\t Reward: -26.697000000000113\n",
      "Episode: 683 \t\t Reward: -26.733000000000114\n",
      "Episode: 684 \t\t Reward: -26.734500000000118\n",
      "Episode: 685 \t\t Reward: -26.733000000000114\n",
      "Episode: 686 \t\t Reward: -26.733000000000114\n",
      "Episode: 687 \t\t Reward: -26.733000000000114\n",
      "Episode: 688 \t\t Reward: -26.731500000000114\n",
      "Episode: 689 \t\t Reward: -26.731500000000114\n",
      "Episode: 690 \t\t Reward: -26.730000000000114\n",
      "Episode: 691 \t\t Reward: -26.73000000000011\n",
      "Episode: 692 \t\t Reward: -26.730000000000114\n",
      "Episode: 693 \t\t Reward: -26.730000000000114\n",
      "Episode: 694 \t\t Reward: -26.450000000000113\n",
      "Episode: 695 \t\t Reward: -26.45000000000011\n",
      "Episode: 696 \t\t Reward: -26.450000000000113\n",
      "Episode: 697 \t\t Reward: -26.448500000000113\n",
      "Episode: 698 \t\t Reward: -26.448500000000113\n",
      "Episode: 699 \t\t Reward: -26.448500000000113\n",
      "Episode: 700 \t\t Reward: -26.447000000000113\n",
      "Episode: 701 \t\t Reward: -26.44550000000011\n",
      "Episode: 702 \t\t Reward: -26.44550000000011\n",
      "Episode: 703 \t\t Reward: -26.44550000000011\n",
      "Episode: 704 \t\t Reward: -26.44550000000011\n",
      "Episode: 705 \t\t Reward: -26.44550000000011\n",
      "Episode: 706 \t\t Reward: -26.447000000000113\n",
      "Episode: 707 \t\t Reward: -26.167000000000115\n",
      "Episode: 708 \t\t Reward: -26.167000000000115\n",
      "Episode: 709 \t\t Reward: -26.16700000000011\n",
      "Episode: 710 \t\t Reward: -26.447000000000113\n",
      "Episode: 711 \t\t Reward: -26.447000000000113\n",
      "Episode: 712 \t\t Reward: -26.447000000000116\n",
      "Episode: 713 \t\t Reward: -26.447000000000113\n",
      "Episode: 714 \t\t Reward: -26.447000000000113\n",
      "Episode: 715 \t\t Reward: -26.448500000000113\n",
      "Episode: 716 \t\t Reward: -26.447000000000113\n",
      "Episode: 717 \t\t Reward: -26.447000000000113\n",
      "Episode: 718 \t\t Reward: -26.447000000000113\n",
      "Episode: 719 \t\t Reward: -26.447000000000113\n",
      "Episode: 720 \t\t Reward: -26.44550000000011\n",
      "Episode: 721 \t\t Reward: -26.44550000000011\n",
      "Episode: 722 \t\t Reward: -26.44550000000011\n",
      "Episode: 723 \t\t Reward: -26.444000000000116\n",
      "Episode: 724 \t\t Reward: -26.444000000000116\n",
      "Episode: 725 \t\t Reward: -26.405000000000108\n",
      "Episode: 726 \t\t Reward: -26.405000000000108\n",
      "Episode: 727 \t\t Reward: -26.40350000000011\n",
      "Episode: 728 \t\t Reward: -26.40350000000011\n",
      "Episode: 729 \t\t Reward: -26.40200000000011\n",
      "Episode: 730 \t\t Reward: -26.402000000000108\n",
      "Episode: 731 \t\t Reward: -26.40200000000011\n",
      "Episode: 732 \t\t Reward: -26.40050000000011\n",
      "Episode: 733 \t\t Reward: -26.40050000000011\n",
      "Episode: 734 \t\t Reward: -26.12050000000011\n",
      "Episode: 735 \t\t Reward: -26.12050000000011\n",
      "Episode: 736 \t\t Reward: -26.119000000000113\n",
      "Episode: 737 \t\t Reward: -26.119000000000113\n",
      "Episode: 738 \t\t Reward: -26.12050000000011\n",
      "Episode: 739 \t\t Reward: -26.12050000000011\n",
      "Episode: 740 \t\t Reward: -26.120500000000117\n",
      "Episode: 741 \t\t Reward: -26.12050000000011\n",
      "Episode: 742 \t\t Reward: -26.11900000000011\n",
      "Episode: 743 \t\t Reward: -26.12050000000011\n",
      "Episode: 744 \t\t Reward: -26.11900000000011\n",
      "Episode: 745 \t\t Reward: -26.11900000000011\n",
      "Episode: 746 \t\t Reward: -26.11900000000011\n",
      "Episode: 747 \t\t Reward: -26.15050000000011\n",
      "Episode: 748 \t\t Reward: -26.15050000000011\n",
      "Episode: 749 \t\t Reward: -26.149000000000115\n",
      "Episode: 750 \t\t Reward: -26.14900000000011\n",
      "Episode: 751 \t\t Reward: -26.14900000000011\n",
      "Episode: 752 \t\t Reward: -26.147500000000107\n",
      "Episode: 753 \t\t Reward: -26.147500000000107\n",
      "Episode: 754 \t\t Reward: -26.146000000000114\n",
      "Episode: 755 \t\t Reward: -26.146000000000114\n",
      "Episode: 756 \t\t Reward: -26.147500000000115\n",
      "Episode: 757 \t\t Reward: -26.147500000000115\n",
      "Episode: 758 \t\t Reward: -26.147500000000115\n",
      "Episode: 759 \t\t Reward: -26.147500000000115\n",
      "Episode: 760 \t\t Reward: -26.146000000000114\n",
      "Episode: 761 \t\t Reward: -26.108500000000113\n",
      "Episode: 762 \t\t Reward: -26.108500000000113\n",
      "Episode: 763 \t\t Reward: -26.108500000000113\n",
      "Episode: 764 \t\t Reward: -26.108500000000113\n",
      "Episode: 765 \t\t Reward: -26.11000000000011\n",
      "Episode: 766 \t\t Reward: -26.110000000000113\n",
      "Episode: 767 \t\t Reward: -26.110000000000113\n",
      "Episode: 768 \t\t Reward: -25.828500000000112\n",
      "Episode: 769 \t\t Reward: -25.82850000000011\n",
      "Episode: 770 \t\t Reward: -25.828500000000112\n",
      "Episode: 771 \t\t Reward: -25.82700000000011\n",
      "Episode: 772 \t\t Reward: -25.82700000000011\n",
      "Episode: 773 \t\t Reward: -25.82700000000011\n",
      "Episode: 774 \t\t Reward: -25.82700000000011\n",
      "Episode: 775 \t\t Reward: -25.82850000000011\n",
      "Episode: 776 \t\t Reward: -26.11000000000011\n",
      "Episode: 777 \t\t Reward: -26.11000000000011\n",
      "Episode: 778 \t\t Reward: -26.390000000000114\n",
      "Episode: 779 \t\t Reward: -26.390000000000114\n",
      "Episode: 780 \t\t Reward: -26.39000000000011\n",
      "Episode: 781 \t\t Reward: -26.39000000000011\n",
      "Episode: 782 \t\t Reward: -26.67000000000011\n",
      "Episode: 783 \t\t Reward: -26.67000000000011\n",
      "Episode: 784 \t\t Reward: -26.668500000000112\n",
      "Episode: 785 \t\t Reward: -26.668500000000112\n",
      "Episode: 786 \t\t Reward: -26.668500000000112\n",
      "Episode: 787 \t\t Reward: -26.668500000000112\n",
      "Episode: 788 \t\t Reward: -26.668500000000112\n",
      "Episode: 789 \t\t Reward: -26.670000000000115\n",
      "Episode: 790 \t\t Reward: -26.670000000000115\n",
      "Episode: 791 \t\t Reward: -26.670000000000115\n",
      "Episode: 792 \t\t Reward: -26.670000000000115\n",
      "Episode: 793 \t\t Reward: -26.67000000000011\n",
      "Episode: 794 \t\t Reward: -26.95000000000011\n",
      "Episode: 795 \t\t Reward: -26.95150000000011\n",
      "Episode: 796 \t\t Reward: -26.95150000000011\n",
      "Episode: 797 \t\t Reward: -26.95150000000011\n",
      "Episode: 798 \t\t Reward: -26.951500000000113\n",
      "Episode: 799 \t\t Reward: -26.951500000000113\n",
      "Episode: 800 \t\t Reward: -26.951500000000113\n",
      "Episode: 801 \t\t Reward: -26.951500000000113\n",
      "Episode: 802 \t\t Reward: -26.951500000000113\n",
      "Episode: 803 \t\t Reward: -26.951500000000113\n",
      "Episode: 804 \t\t Reward: -26.951500000000113\n",
      "Episode: 805 \t\t Reward: -26.950000000000113\n",
      "Episode: 806 \t\t Reward: -26.948500000000113\n",
      "Episode: 807 \t\t Reward: -27.228500000000114\n",
      "Episode: 808 \t\t Reward: -27.228500000000114\n",
      "Episode: 809 \t\t Reward: -27.228500000000114\n",
      "Episode: 810 \t\t Reward: -27.228500000000114\n",
      "Episode: 811 \t\t Reward: -27.228500000000114\n",
      "Episode: 812 \t\t Reward: -27.230000000000114\n",
      "Episode: 813 \t\t Reward: -27.230000000000114\n",
      "Episode: 814 \t\t Reward: -26.950000000000113\n",
      "Episode: 815 \t\t Reward: -26.948500000000116\n",
      "Episode: 816 \t\t Reward: -26.948500000000113\n",
      "Episode: 817 \t\t Reward: -26.948500000000116\n",
      "Episode: 818 \t\t Reward: -26.948500000000113\n",
      "Episode: 819 \t\t Reward: -26.948500000000113\n",
      "Episode: 820 \t\t Reward: -26.948500000000113\n",
      "Episode: 821 \t\t Reward: -26.948500000000113\n",
      "Episode: 822 \t\t Reward: -26.948500000000113\n",
      "Episode: 823 \t\t Reward: -26.948500000000113\n",
      "Episode: 824 \t\t Reward: -26.948500000000113\n",
      "Episode: 825 \t\t Reward: -26.986000000000114\n",
      "Episode: 826 \t\t Reward: -26.986000000000114\n",
      "Episode: 827 \t\t Reward: -26.987500000000114\n",
      "Episode: 828 \t\t Reward: -26.987500000000114\n",
      "Episode: 829 \t\t Reward: -26.989000000000114\n",
      "Episode: 830 \t\t Reward: -26.989000000000114\n",
      "Episode: 831 \t\t Reward: -26.989000000000114\n",
      "Episode: 832 \t\t Reward: -26.990500000000115\n",
      "Episode: 833 \t\t Reward: -26.710500000000117\n",
      "Episode: 834 \t\t Reward: -26.990500000000115\n",
      "Episode: 835 \t\t Reward: -26.990500000000115\n",
      "Episode: 836 \t\t Reward: -26.990500000000115\n",
      "Episode: 837 \t\t Reward: -26.990500000000115\n",
      "Episode: 838 \t\t Reward: -26.989000000000114\n",
      "Episode: 839 \t\t Reward: -26.989000000000114\n",
      "Episode: 840 \t\t Reward: -26.990500000000115\n",
      "Episode: 841 \t\t Reward: -26.990500000000115\n",
      "Episode: 842 \t\t Reward: -26.990500000000115\n",
      "Episode: 843 \t\t Reward: -26.989000000000118\n",
      "Episode: 844 \t\t Reward: -26.989000000000114\n",
      "Episode: 845 \t\t Reward: -26.989000000000114\n",
      "Episode: 846 \t\t Reward: -26.989000000000114\n",
      "Episode: 847 \t\t Reward: -26.989000000000114\n",
      "Episode: 848 \t\t Reward: -26.990500000000115\n",
      "Episode: 849 \t\t Reward: -26.990500000000115\n",
      "Episode: 850 \t\t Reward: -26.990500000000115\n",
      "Episode: 851 \t\t Reward: -26.990500000000115\n",
      "Episode: 852 \t\t Reward: -26.710500000000117\n",
      "Episode: 853 \t\t Reward: -26.710500000000117\n",
      "Episode: 854 \t\t Reward: -26.71050000000011\n",
      "Episode: 855 \t\t Reward: -26.71050000000011\n",
      "Episode: 856 \t\t Reward: -26.709000000000113\n",
      "Episode: 857 \t\t Reward: -26.710500000000117\n",
      "Episode: 858 \t\t Reward: -26.710500000000117\n",
      "Episode: 859 \t\t Reward: -26.712000000000117\n",
      "Episode: 860 \t\t Reward: -26.712000000000117\n",
      "Episode: 861 \t\t Reward: -26.749500000000115\n",
      "Episode: 862 \t\t Reward: -26.74950000000011\n",
      "Episode: 863 \t\t Reward: -26.749500000000115\n",
      "Episode: 864 \t\t Reward: -26.749500000000115\n",
      "Episode: 865 \t\t Reward: -26.748000000000115\n",
      "Episode: 866 \t\t Reward: -26.74950000000011\n",
      "Episode: 867 \t\t Reward: -26.75100000000012\n",
      "Episode: 868 \t\t Reward: -27.031000000000116\n",
      "Episode: 869 \t\t Reward: -27.03250000000012\n",
      "Episode: 870 \t\t Reward: -27.03250000000012\n",
      "Episode: 871 \t\t Reward: -27.032500000000113\n",
      "Episode: 872 \t\t Reward: -27.032500000000113\n",
      "Episode: 873 \t\t Reward: -27.032500000000113\n",
      "Episode: 874 \t\t Reward: -27.032500000000113\n",
      "Episode: 875 \t\t Reward: -27.031000000000112\n",
      "Episode: 876 \t\t Reward: -27.029500000000116\n",
      "Episode: 877 \t\t Reward: -27.029500000000116\n",
      "Episode: 878 \t\t Reward: -27.031000000000116\n",
      "Episode: 879 \t\t Reward: -27.031000000000116\n",
      "Episode: 880 \t\t Reward: -27.031000000000116\n",
      "Episode: 881 \t\t Reward: -27.031000000000116\n",
      "Episode: 882 \t\t Reward: -27.032500000000113\n",
      "Episode: 883 \t\t Reward: -27.032500000000113\n",
      "Episode: 884 \t\t Reward: -27.032500000000113\n",
      "Episode: 885 \t\t Reward: -27.032500000000113\n",
      "Episode: 886 \t\t Reward: -27.032500000000113\n",
      "Episode: 887 \t\t Reward: -27.03250000000012\n",
      "Episode: 888 \t\t Reward: -27.03250000000012\n",
      "Episode: 889 \t\t Reward: -27.031000000000112\n",
      "Episode: 890 \t\t Reward: -26.75100000000011\n",
      "Episode: 891 \t\t Reward: -26.75100000000011\n",
      "Episode: 892 \t\t Reward: -26.75100000000011\n",
      "Episode: 893 \t\t Reward: -26.471000000000114\n",
      "Episode: 894 \t\t Reward: -26.472500000000114\n",
      "Episode: 895 \t\t Reward: -26.471000000000117\n",
      "Episode: 896 \t\t Reward: -26.471000000000117\n",
      "Episode: 897 \t\t Reward: -26.471000000000117\n",
      "Episode: 898 \t\t Reward: -26.471000000000114\n",
      "Episode: 899 \t\t Reward: -26.471000000000114\n",
      "Episode: 900 \t\t Reward: -26.472500000000114\n",
      "Episode: 901 \t\t Reward: -26.472500000000114\n",
      "Episode: 902 \t\t Reward: -26.471000000000117\n",
      "Episode: 903 \t\t Reward: -26.472500000000117\n",
      "Episode: 904 \t\t Reward: -26.472500000000114\n",
      "Episode: 905 \t\t Reward: -26.472500000000114\n",
      "Episode: 906 \t\t Reward: -26.472500000000114\n",
      "Episode: 907 \t\t Reward: -26.472500000000114\n",
      "Episode: 908 \t\t Reward: -26.47250000000011\n",
      "Episode: 909 \t\t Reward: -26.47250000000011\n",
      "Episode: 910 \t\t Reward: -26.47250000000011\n",
      "Episode: 911 \t\t Reward: -26.47400000000011\n",
      "Episode: 912 \t\t Reward: -26.472500000000114\n",
      "Episode: 913 \t\t Reward: -26.47250000000011\n",
      "Episode: 914 \t\t Reward: -26.752500000000115\n",
      "Episode: 915 \t\t Reward: -26.752500000000108\n",
      "Episode: 916 \t\t Reward: -26.752500000000115\n",
      "Episode: 917 \t\t Reward: -26.752500000000108\n",
      "Episode: 918 \t\t Reward: -26.752500000000115\n",
      "Episode: 919 \t\t Reward: -26.752500000000115\n",
      "Episode: 920 \t\t Reward: -26.752500000000115\n",
      "Episode: 921 \t\t Reward: -26.472500000000114\n",
      "Episode: 922 \t\t Reward: -26.474000000000114\n",
      "Episode: 923 \t\t Reward: -26.474000000000114\n",
      "Episode: 924 \t\t Reward: -26.194000000000116\n",
      "Episode: 925 \t\t Reward: -26.194000000000116\n",
      "Episode: 926 \t\t Reward: -26.194000000000116\n",
      "Episode: 927 \t\t Reward: -26.192500000000113\n",
      "Episode: 928 \t\t Reward: -26.19250000000011\n",
      "Episode: 929 \t\t Reward: -26.191000000000113\n",
      "Episode: 930 \t\t Reward: -26.191000000000113\n",
      "Episode: 931 \t\t Reward: -26.191000000000113\n",
      "Episode: 932 \t\t Reward: -26.191000000000116\n",
      "Episode: 933 \t\t Reward: -26.472500000000117\n",
      "Episode: 934 \t\t Reward: -26.472500000000114\n",
      "Episode: 935 \t\t Reward: -26.472500000000114\n",
      "Episode: 936 \t\t Reward: -26.474000000000114\n",
      "Episode: 937 \t\t Reward: -26.474000000000114\n",
      "Episode: 938 \t\t Reward: -26.474000000000114\n",
      "Episode: 939 \t\t Reward: -26.474000000000114\n",
      "Episode: 940 \t\t Reward: -26.47250000000011\n",
      "Episode: 941 \t\t Reward: -26.47250000000011\n",
      "Episode: 942 \t\t Reward: -26.472500000000114\n",
      "Episode: 943 \t\t Reward: -26.192500000000113\n",
      "Episode: 944 \t\t Reward: -26.19400000000011\n",
      "Episode: 945 \t\t Reward: -26.194000000000116\n",
      "Episode: 946 \t\t Reward: -26.194000000000116\n",
      "Episode: 947 \t\t Reward: -26.19400000000011\n",
      "Episode: 948 \t\t Reward: -26.19250000000011\n",
      "Episode: 949 \t\t Reward: -26.19250000000011\n",
      "Episode: 950 \t\t Reward: -26.19250000000011\n",
      "Episode: 951 \t\t Reward: -26.19250000000011\n",
      "Episode: 952 \t\t Reward: -26.47250000000011\n",
      "Episode: 953 \t\t Reward: -26.472500000000114\n",
      "Episode: 954 \t\t Reward: -26.47250000000011\n",
      "Episode: 955 \t\t Reward: -26.472500000000114\n",
      "Episode: 956 \t\t Reward: -26.47250000000011\n",
      "Episode: 957 \t\t Reward: -26.472500000000117\n",
      "Episode: 958 \t\t Reward: -26.192500000000113\n",
      "Episode: 959 \t\t Reward: -26.191000000000116\n",
      "Episode: 960 \t\t Reward: -26.192500000000113\n",
      "Episode: 961 \t\t Reward: -26.192500000000113\n",
      "Episode: 962 \t\t Reward: -26.192500000000113\n",
      "Episode: 963 \t\t Reward: -26.192500000000113\n",
      "Episode: 964 \t\t Reward: -26.19250000000011\n",
      "Episode: 965 \t\t Reward: -26.19250000000011\n",
      "Episode: 966 \t\t Reward: -26.191000000000113\n",
      "Episode: 967 \t\t Reward: -26.189500000000113\n",
      "Episode: 968 \t\t Reward: -26.18950000000011\n",
      "Episode: 969 \t\t Reward: -26.188000000000113\n",
      "Episode: 970 \t\t Reward: -26.188000000000113\n",
      "Episode: 971 \t\t Reward: -26.188000000000113\n",
      "Episode: 972 \t\t Reward: -26.188000000000113\n",
      "Episode: 973 \t\t Reward: -26.189500000000113\n",
      "Episode: 974 \t\t Reward: -26.189500000000113\n",
      "Episode: 975 \t\t Reward: -26.189500000000113\n",
      "Episode: 976 \t\t Reward: -26.189500000000113\n",
      "Episode: 977 \t\t Reward: -26.189500000000113\n",
      "Episode: 978 \t\t Reward: -26.188000000000113\n",
      "Episode: 979 \t\t Reward: -26.188000000000113\n",
      "Episode: 980 \t\t Reward: -26.189500000000113\n",
      "Episode: 981 \t\t Reward: -26.189500000000113\n",
      "Episode: 982 \t\t Reward: -26.189500000000116\n",
      "Episode: 983 \t\t Reward: -26.189500000000116\n",
      "Episode: 984 \t\t Reward: -26.189500000000116\n",
      "Episode: 985 \t\t Reward: -26.191000000000113\n",
      "Episode: 986 \t\t Reward: -26.192500000000113\n",
      "Episode: 987 \t\t Reward: -26.192500000000113\n",
      "Episode: 988 \t\t Reward: -26.194000000000116\n",
      "Episode: 989 \t\t Reward: -26.194000000000116\n",
      "Episode: 990 \t\t Reward: -26.474000000000114\n",
      "Episode: 991 \t\t Reward: -26.474000000000114\n",
      "Episode: 992 \t\t Reward: -26.47400000000011\n",
      "Episode: 993 \t\t Reward: -26.754000000000115\n",
      "Episode: 994 \t\t Reward: -26.752500000000115\n",
      "Episode: 995 \t\t Reward: -26.472500000000114\n",
      "Episode: 996 \t\t Reward: -26.472500000000114\n",
      "Episode: 997 \t\t Reward: -26.472500000000114\n",
      "Episode: 998 \t\t Reward: -26.192500000000113\n",
      "Episode: 999 \t\t Reward: -26.19400000000011\n",
      "Episode: 1000 \t\t Reward: -26.192500000000113\n",
      "Total test time: 03:06:43.41\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from collections import deque\n",
    "import numpy as np\n",
    "from utility import get_unique_numbers\n",
    "import time\n",
    "\n",
    "test_running_reward = 0\n",
    "\n",
    "ac =[]\n",
    "start = time.time()\n",
    "runn_ep =[]\n",
    "scores =[]\n",
    "scores_deque = deque(maxlen=100)\n",
    "avg_scores =[]\n",
    "Bmp = []\n",
    "T_Bmp = []\n",
    "bumps = 0\n",
    "\n",
    "if p_model.is_dir():\n",
    "    nb_name = 'DQN_Test_log.txt'\n",
    "with open(p_model.joinpath(nb_name), \"w\") as f:\n",
    "\n",
    "    for ep in range(1, total_test_episodes+1):\n",
    "        start_ep = time.time()\n",
    "        ep_reward = 0\n",
    "        runn_ep.append(ep)\n",
    "        state = env.reset()\n",
    "        Ep_bmp = []\n",
    "\n",
    "        for t in range(1, 1500):\n",
    "            action = agent_DQN.act(state)\n",
    "            ac.append(action)\n",
    "            state, reward, done, bumps = env.step(action)\n",
    "\n",
    "            Ep_bmp.append(bumps)\n",
    "            T_Bmp.append(bumps)\n",
    "            ep_reward += reward\n",
    "            # env.render('rgb_array') # uncommand to display -- warning: expensive resources.\n",
    "            if done:\n",
    "                break\n",
    "        \n",
    "        Bmp.append(Ep_bmp)\n",
    "        scores_deque.append(ep_reward)\n",
    "        avg_scores.append(np.mean(scores_deque))\n",
    "        end_ep = time.time()\n",
    "        hours_ep, rem_ep = divmod(end_ep-start_ep, 3600)\n",
    "        minutes_ep, seconds_ep = divmod(rem_ep, 60)\n",
    "\n",
    "        f.write(\"Episode{} (survive time): {:0>2}:{:0>2}:{:05.2f}\\n\".format(ep,int(hours_ep),int(minutes_ep),seconds_ep))\n",
    "        test_running_reward +=  ep_reward\n",
    "\n",
    "        print('Episode: {} \\t\\t Reward: {}'.format(ep, avg_scores[ep-1]))\n",
    "\n",
    "        scores.append(round(ep_reward, 2))\n",
    "        ep_reward = 0\n",
    "    env.close()\n",
    "\n",
    "    end = time.time()\n",
    "    hours, rem = divmod(end-start, 3600)\n",
    "    minutes, seconds = divmod(rem, 60)\n",
    "    print(\"Total test time: {:0>2}:{:0>2}:{:05.2f}\".format(int(hours),int(minutes),seconds))\n",
    "    \n",
    "    # count the bump instances.\n",
    "    count_list_DQN = []\n",
    "    for i in range(len(Bmp)):\n",
    "        count_list_DQN.append(count_impulses(Bmp[i]))\n",
    "    data_dqn = {'Bump_instance': count_list_DQN}\n",
    "    Bump_track = pd.DataFrame(data_dqn)\n",
    "    Tcount_DQN = count_edgezone(T_Bmp,gap_threshold)\n",
    "\n",
    "    f.write(\"Total Test time (survive time): {:0>2}:{:0>2}:{:05.2f}\\n\".format(int(hours),int(minutes),seconds))\n",
    "    f.write('Toal epoch:%s\\n'%total_test_episodes)\n",
    "    f.write(\"Bump instances of total: %s\\n\"%Tcount_DQN)\n",
    "    f.write(\"Bump to edge zone %s\\n\"%count_edgezone(T_Bmp, gap_threshold))\n",
    "    f.close()\n",
    "\n",
    "DQN_avg = avg_scores\n",
    "DQN_scores = scores\n",
    "DQN_ac = ac\n",
    "\n",
    "csv_name = f\"DQN_reshape\"\n",
    "# score_save(scores,avg_scores,file_name=csv_name,path=p_model,print_=False) # save score table as csv file\n",
    "score_action_save(ac,scores,avg_scores,file_name=csv_name,path=p_model,print_=False) # save score table as csv file\n",
    "count_impulse_file = f\"DQN_reshape_count\"\n",
    "Bump_track.to_csv((p_model/f\"{count_impulse_file}.csv\")) # save bump instances as csv file\n",
    "\n",
    "\n",
    "fig = plt.figure()\n",
    "ax = fig.add_subplot(111)\n",
    "plt.plot(avg_scores)\n",
    "# plt.plot(df['Averagescores'])\n",
    "plt.ylabel('Scores')\n",
    "plt.xlabel(f'Episode {len(avg_scores)}')\n",
    "plt.show()\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "\n",
    "action_list = [0,1,2,3,4]\n",
    "h=[]\n",
    "for i in action_list:\n",
    "    h.append(ac.count(i))\n",
    "bar_contain = plt.bar(action_list,h)\n",
    "plt.xlabel('Action items')\n",
    "plt.ylabel('Number of action item used during trial')\n",
    "plt.bar_label(bar_contain)\n",
    "plt.show()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Reshape plot\n",
    "## Score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "DDPG_ac = pd.Series(DDPG_ac, name='DDPG_ac')\n",
    "DDPG_scores = pd.Series(DDPG_scores, name='DDPG_scores')\n",
    "DDPG_average_score=pd.Series(DDPG_avg,name='DDPG_avg')\n",
    "\n",
    "PPO_ac = pd.Series(PPO_ac, name='PPO_ac')\n",
    "PPO_scores = pd.Series(PPO_scores, name='PPO_scores')\n",
    "PPO_average_score=pd.Series(PPO_avg,name='PPO_avg')\n",
    "\n",
    "DQN_ac = pd.Series(DQN_ac, name='DQN_ac')\n",
    "DQN_scores = pd.Series(DQN_scores, name='DQN_scores')\n",
    "DQN_average_score=pd.Series(DQN_avg,name='DQN_avg')\n",
    "\n",
    "\n",
    "df = pd.concat([DQN_ac,DQN_scores, DQN_average_score, DDPG_ac,DDPG_scores,DDPG_average_score,PPO_ac,PPO_scores,PPO_average_score],axis=1)\n",
    "\n",
    "fig = plt.figure()\n",
    "ax = fig.add_subplot(111)\n",
    "plt.plot(df['DQN_avg'], 'r-.',label='DQN Average Score')\n",
    "plt.plot(df['DDPG_avg'],'b',label='DDPG Average Score')\n",
    "plt.plot(df['PPO_avg'],'k--' ,label='PPO Average Score')\n",
    "plt.ylim([-650, 100])\n",
    "# plt.plot(df['Averagescores'])\n",
    "plt.ylabel('Average Scores')\n",
    "plt.xlabel('Episode')\n",
    "plt.legend()\n",
    "# plt.xlabel(f'Episode {len(df['DQN_avg'])}')\n",
    "\n",
    "plt.savefig((p_model/'Test_average2.png'))\n",
    "plt.show()\n",
    "df.to_csv((p_model/'scores.csv'))"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Actions plan"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "action_list = [0,1,2,3,4]\n",
    "action_item = ('Do nothing', 'Turn left', 'Turn right','Accelerate','Brake')\n",
    "h1= []\n",
    "h2 =[]\n",
    "h3 =[]\n",
    "\n",
    "for i in action_list:\n",
    "    h1.append(df[df['PPO_ac']==i]['PPO_ac'].count())\n",
    "    h2.append(df[df['DDPG_ac']==i]['DDPG_ac'].count())\n",
    "    h3.append(df[df['DQN_ac']==i]['DQN_ac'].count())\n",
    "x_labels = np.array(action_list)\n",
    "bar_width = 0.35\n",
    "\n",
    "fig, ax = plt.subplots()\n",
    "\n",
    "ax.bar(x_labels - bar_width/2, h1, bar_width, label='PPO agent',color='b')\n",
    "ax.bar(x_labels + bar_width/2, h2, bar_width, label='DDPG agent',color='y')\n",
    "ax.bar(x_labels - bar_width-0.2/2, h3, bar_width, label='DQN agent',color='k')\n",
    "\n",
    "ax.set_xlabel('Action')\n",
    "ax.set_ylabel('Count')\n",
    "ax.set_title('Counts of Actions')\n",
    "plt.legend('lower center')\n",
    "plt.xticks(x_labels,action_item)\n",
    "\n",
    "plt.savefig((p_model/'Test_action_counts2.png'))\n",
    "\n",
    "plt.show()\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "RLmac",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.15"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}