-
Notifications
You must be signed in to change notification settings - Fork 4
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
1 parent
2688947
commit 5f09d2f
Showing
68 changed files
with
285 additions
and
0 deletions.
There are no files selected for viewing
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,25 @@ | ||
DQN Architecture and Hyperparameters: | ||
DQNAgent: | ||
- number of agents: 2 | ||
- state_size: self.v_field_res+ 1, action_size=3 | ||
- action_size: 3 [1: explore, 2: exploit, 3: relocate] | ||
- replay_memory_capacity: 10,000 | ||
- batch_size: 128 | ||
- gamma: 0.99 | ||
- epsilon_start: 0.9 | ||
- tau: 0.005 | ||
- epsilon_decay: 1000 | ||
- epsilon_end: 0.05 | ||
- lr: 1e-4 | ||
- reward = collected / time | ||
|
||
DQNetwork: | ||
- input_size: [Specify the size of the input] | ||
- output_size: [Specify the size of the output layer] | ||
|
||
Training Process: | ||
- Experience replay with a deque (max capacity: 10,000) | ||
- Epsilon-greedy exploration | ||
- Q-network trained with mini-batches (batch size: 128) | ||
- Mean Squared Error (MSE) loss | ||
- Target Q-network updated with soft update (tau: 0.005) every step |
Binary file not shown.
Binary file not shown.
Binary file added
BIN
+30.5 KB
...nt/experiments/exp_0/tf_logs/events.out.tfevents.1704280575.Feriels-MBP.fritz.box.31020.0
Binary file not shown.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,25 @@ | ||
DQN Architecture and Hyperparameters: | ||
DQNAgent: | ||
- number of agents: 2 | ||
- state_size: self.v_field_res+ 1, action_size=3 | ||
- action_size: 3 [1: explore, 2: exploit, 3: relocate] | ||
- replay_memory_capacity: 10,000 | ||
- batch_size: 128 | ||
- gamma: 0.99 | ||
- epsilon_start: 0.9 | ||
- tau: 0.005 | ||
- epsilon_decay: 1000 | ||
- epsilon_end: 0.05 | ||
- lr: 1e-5 | ||
- reward = collected / time | ||
|
||
DQNetwork: | ||
- input_size: [Specify the size of the input] | ||
- output_size: [Specify the size of the output layer] | ||
|
||
Training Process: | ||
- Experience replay with a deque (max capacity: 10,000) | ||
- Epsilon-greedy exploration | ||
- Q-network trained with mini-batches (batch size: 128) | ||
- Mean Squared Error (MSE) loss | ||
- Target Q-network updated with soft update (tau: 0.005) every step |
Binary file not shown.
Binary file not shown.
Binary file added
BIN
+30.5 KB
...nt/experiments/exp_1/tf_logs/events.out.tfevents.1704285608.Feriels-MBP.fritz.box.31812.0
Binary file not shown.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,27 @@ | ||
DQN-Architecture and Hyperparameters: | ||
|
||
DQNAgent: | ||
- number of agents: 3 | ||
- state_size: self.v_field_res+ 1, action_size=3 | ||
- action_size: 3 [1: explore, 2: exploit, 3: relocate] | ||
- replay_memory_capacity: 10,000 | ||
- batch_size: 128 | ||
- gamma: 0.99 | ||
- epsilon_start: 0.9 | ||
- tau: 0.005 | ||
- epsilon_decay: 1000 | ||
- epsilon_end: 0.05 | ||
- lr: 1e-5 with scheduler | ||
- if self.t!=0 reward= (0.2*ag.collected_r + 0.8*collective_reward) /self.t else reward=0 | ||
where collective_reward = sum of ag.collected_r / (self.t*len(agents)) | ||
|
||
DQNetwork: | ||
- input_size: [Specify the size of the input] | ||
- output_size: [Specify the size of the output layer] | ||
|
||
Training Process: | ||
- Experience replay with a deque (max capacity: 10,000) | ||
- Epsilon-greedy exploration | ||
- Q-network trained with mini-batches (batch size: 128) | ||
- Mean Squared Error (MSE) loss | ||
- Target Q-network updated with soft update (tau: 0.005) every 50 iterations |
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file added
BIN
+92.4 KB
...periments/exp_10/tf_logs/events.out.tfevents.1704708244.Feriels-MacBook-Pro.local.57757.0
Binary file not shown.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,26 @@ | ||
DQN Architecture and Hyperparameters: | ||
DQNAgent: | ||
- number of agents: 1 | ||
- state_size: self.v_field_res+ 1, action_size=3 | ||
- action_size: 3 [1: explore, 2: exploit, 3: relocate] | ||
- replay_memory_capacity: 10,000 | ||
- batch_size: 128 | ||
- gamma: 0.99 | ||
- epsilon_start: 0.9 | ||
- tau: 0.005 | ||
- epsilon_decay: 1000 | ||
- epsilon_end: 0.05 | ||
- lr: 1e-5 with scheduler | ||
- reward = collected / time | ||
|
||
|
||
DQNetwork: | ||
- input_size: [Specify the size of the input] | ||
- output_size: [Specify the size of the output layer] | ||
|
||
Training Process: | ||
- Experience replay with a deque (max capacity: 10,000) | ||
- Epsilon-greedy exploration | ||
- Q-network trained with mini-batches (batch size: 128) | ||
- Mean Squared Error (MSE) loss | ||
- Target Q-network updated with soft update (tau: 0.005) every step |
Binary file not shown.
Binary file added
BIN
+15.3 KB
...nt/experiments/exp_2/tf_logs/events.out.tfevents.1704297112.Feriels-MBP.fritz.box.32628.0
Binary file not shown.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,26 @@ | ||
DQN Architecture and Hyperparameters: | ||
DQNAgent: | ||
- number of agents: 1 | ||
- state_size: self.v_field_res+ 1, action_size=3 | ||
- action_size: 3 [1: explore, 2: exploit, 3: relocate] | ||
- replay_memory_capacity: 10,000 | ||
- batch_size: 128 | ||
- gamma: 0.99 | ||
- epsilon_start: 0.9 | ||
- tau: 0.005 | ||
- epsilon_decay: 1000 | ||
- epsilon_end: 0.05 | ||
- lr: 1e-5 with scheduler | ||
- reward = collected / time | ||
|
||
|
||
DQNetwork: | ||
- input_size: [Specify the size of the input] | ||
- output_size: [Specify the size of the output layer] | ||
|
||
Training Process: | ||
- Experience replay with a deque (max capacity: 10,000) | ||
- Epsilon-greedy exploration | ||
- Q-network trained with mini-batches (batch size: 128) | ||
- Mean Squared Error (MSE) loss | ||
- Target Q-network updated with soft update (tau: 0.005) every 50 iterations |
Binary file not shown.
Binary file added
BIN
+88 Bytes
...nt/experiments/exp_3/tf_logs/events.out.tfevents.1704302074.Feriels-MBP.fritz.box.33609.0
Binary file not shown.
Binary file added
BIN
+88 Bytes
...nt/experiments/exp_3/tf_logs/events.out.tfevents.1704302118.Feriels-MBP.fritz.box.33626.0
Binary file not shown.
Binary file added
BIN
+3.39 KB
...nt/experiments/exp_3/tf_logs/events.out.tfevents.1704303586.Feriels-MBP.fritz.box.33871.0
Binary file not shown.
Binary file added
BIN
+15.3 KB
...nt/experiments/exp_3/tf_logs/events.out.tfevents.1704303788.Feriels-MBP.fritz.box.33909.0
Binary file not shown.
Binary file added
BIN
+15.3 KB
...nt/experiments/exp_3/tf_logs/events.out.tfevents.1704309517.Feriels-MBP.fritz.box.34798.0
Binary file not shown.
Binary file added
BIN
+30.5 KB
...nt/experiments/exp_3/tf_logs/events.out.tfevents.1704310165.Feriels-MBP.fritz.box.34920.0
Binary file not shown.
Binary file added
BIN
+51.1 KB
...nt/experiments/exp_3/tf_logs/events.out.tfevents.1704312280.Feriels-MBP.fritz.box.35263.0
Binary file not shown.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,25 @@ | ||
DQN Architecture and Hyperparameters: | ||
DQNAgent: | ||
- number of agents: 1 | ||
- state_size: self.v_field_res+ 1, action_size=3 | ||
- action_size: 3 [1: explore, 2: exploit, 3: relocate] | ||
- replay_memory_capacity: 10,000 | ||
- batch_size: 128 | ||
- gamma: 0.99 | ||
- epsilon_start: 0.9 | ||
- tau: 0.005 | ||
- epsilon_decay: 1000 | ||
- epsilon_end: 0.05 | ||
- lr: 1e-5 with scheduler | ||
- reward = 1 if resource is exploited, 0 otherwise | ||
|
||
DQNetwork: | ||
- input_size: [Specify the size of the input] | ||
- output_size: [Specify the size of the output layer] | ||
|
||
Training Process: | ||
- Experience replay with a deque (max capacity: 10,000) | ||
- Epsilon-greedy exploration | ||
- Q-network trained with mini-batches (batch size: 128) | ||
- Mean Squared Error (MSE) loss | ||
- Target Q-network updated with soft update (tau: 0.005) every 50 iterations |
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Binary file not shown.
Binary file not shown.
Binary file added
BIN
+51.1 KB
...xperiments/exp_4/tf_logs/events.out.tfevents.1704539912.Feriels-MacBook-Pro.local.46631.0
Binary file not shown.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,26 @@ | ||
DQN Architecture and Hyperparameters: | ||
DQNAgent: | ||
- number of agents: 2 | ||
- state_size: self.v_field_res+ 1, action_size=3 | ||
- action_size: 3 [1: explore, 2: exploit, 3: relocate] | ||
- replay_memory_capacity: 10,000 | ||
- batch_size: 128 | ||
- gamma: 0.99 | ||
- epsilon_start: 0.9 | ||
- tau: 0.005 | ||
- epsilon_decay: 1000 | ||
- epsilon_end: 0.05 | ||
- lr: 1e-5 with scheduler | ||
- o if not exploit else reward = collected / time | ||
|
||
|
||
DQNetwork: | ||
- input_size: [Specify the size of the input] | ||
- output_size: [Specify the size of the output layer] | ||
|
||
Training Process: | ||
- Experience replay with a deque (max capacity: 10,000) | ||
- Epsilon-greedy exploration | ||
- Q-network trained with mini-batches (batch size: 128) | ||
- Mean Squared Error (MSE) loss | ||
- Target Q-network updated with soft update (tau: 0.005) every 50 iterations |
Binary file not shown.
Binary file not shown.
Binary file added
BIN
+51.1 KB
...xperiments/exp_5/tf_logs/events.out.tfevents.1704543170.Feriels-MacBook-Pro.local.47226.0
Binary file not shown.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,26 @@ | ||
DQN-Architecture and Hyperparameters: | ||
|
||
DQNAgent: | ||
- number of agents: 2 | ||
- state_size: self.v_field_res+ 1, action_size=3 | ||
- action_size: 3 [1: explore, 2: exploit, 3: relocate] | ||
- replay_memory_capacity: 10,000 | ||
- batch_size: 128 | ||
- gamma: 0.99 | ||
- epsilon_start: 0.9 | ||
- tau: 0.005 | ||
- epsilon_decay: 1000 | ||
- epsilon_end: 0.05 | ||
- lr: 1e-5 with scheduler | ||
- if self.t!=0 reward= ag.collected_r /self.t else reward=0 | ||
|
||
DQNetwork: | ||
- input_size: [Specify the size of the input] | ||
- output_size: [Specify the size of the output layer] | ||
|
||
Training Process: | ||
- Experience replay with a deque (max capacity: 10,000) | ||
- Epsilon-greedy exploration | ||
- Q-network trained with mini-batches (batch size: 128) | ||
- Mean Squared Error (MSE) loss | ||
- Target Q-network updated with soft update (tau: 0.005) every 50 iterations |
Binary file not shown.
Binary file not shown.
Binary file added
BIN
+88 Bytes
...xperiments/exp_6/tf_logs/events.out.tfevents.1704546260.Feriels-MacBook-Pro.local.47833.0
Binary file not shown.
Binary file added
BIN
+51.1 KB
...xperiments/exp_6/tf_logs/events.out.tfevents.1704546641.Feriels-MacBook-Pro.local.47913.0
Binary file not shown.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,26 @@ | ||
DQN-Architecture and Hyperparameters: | ||
|
||
DQNAgent: | ||
- number of agents: 3 | ||
- state_size: self.v_field_res+ 1, action_size=3 | ||
- action_size: 3 [1: explore, 2: exploit, 3: relocate] | ||
- replay_memory_capacity: 10,000 | ||
- batch_size: 128 | ||
- gamma: 0.99 | ||
- epsilon_start: 0.9 | ||
- tau: 0.005 | ||
- epsilon_decay: 1000 | ||
- epsilon_end: 0.05 | ||
- lr: 1e-5 with scheduler | ||
- if self.t!=0 reward= ag.collected_r /self.t else reward=0 | ||
|
||
DQNetwork: | ||
- input_size: [Specify the size of the input] | ||
- output_size: [Specify the size of the output layer] | ||
|
||
Training Process: | ||
- Experience replay with a deque (max capacity: 10,000) | ||
- Epsilon-greedy exploration | ||
- Q-network trained with mini-batches (batch size: 128) | ||
- Mean Squared Error (MSE) loss | ||
- Target Q-network updated with soft update (tau: 0.005) every 50 iterations |
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file added
BIN
+139 KB
...xperiments/exp_7/tf_logs/events.out.tfevents.1704559485.Feriels-MacBook-Pro.local.49934.0
Binary file not shown.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,26 @@ | ||
DQN-Architecture and Hyperparameters: | ||
|
||
DQNAgent: | ||
- number of agents: 3 | ||
- state_size: self.v_field_res+ 1, action_size=3 | ||
- action_size: 3 [1: explore, 2: exploit, 3: relocate] | ||
- replay_memory_capacity: 10,000 | ||
- batch_size: 128 | ||
- gamma: 0.99 | ||
- epsilon_start: 0.9 | ||
- tau: 0.005 | ||
- epsilon_decay: 1000 | ||
- epsilon_end: 0.05 | ||
- lr: 1e-5 with scheduler | ||
- if self.t!=0 reward= ag.collected_r + ag.collective_reward /self.t else reward=0 | ||
|
||
DQNetwork: | ||
- input_size: [Specify the size of the input] | ||
- output_size: [Specify the size of the output layer] | ||
|
||
Training Process: | ||
- Experience replay with a deque (max capacity: 10,000) | ||
- Epsilon-greedy exploration | ||
- Q-network trained with mini-batches (batch size: 128) | ||
- Mean Squared Error (MSE) loss | ||
- Target Q-network updated with soft update (tau: 0.005) every 50 iterations |
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file added
BIN
+139 KB
...xperiments/exp_8/tf_logs/events.out.tfevents.1704625868.Feriels-MacBook-Pro.local.53954.0
Binary file not shown.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,27 @@ | ||
DQN-Architecture and Hyperparameters: | ||
|
||
DQNAgent: | ||
- number of agents: 3 | ||
- state_size: self.v_field_res+ 1, action_size=3 | ||
- action_size: 3 [1: explore, 2: exploit, 3: relocate] | ||
- replay_memory_capacity: 10,000 | ||
- batch_size: 128 | ||
- gamma: 0.99 | ||
- epsilon_start: 0.9 | ||
- tau: 0.005 | ||
- epsilon_decay: 1000 | ||
- epsilon_end: 0.05 | ||
- lr: 1e-5 with scheduler | ||
- if self.t!=0 reward= (0.2*ag.collected_r + 0.8*collective_reward) /self.t else reward=0 | ||
where collective_reward = sum of ag.collected_r / (self.t*len(agents)) | ||
|
||
DQNetwork: | ||
- input_size: [Specify the size of the input] | ||
- output_size: [Specify the size of the output layer] | ||
|
||
Training Process: | ||
- Experience replay with a deque (max capacity: 10,000) | ||
- Epsilon-greedy exploration | ||
- Q-network trained with mini-batches (batch size: 128) | ||
- Mean Squared Error (MSE) loss | ||
- Target Q-network updated with soft update (tau: 0.005) every 50 iterations |
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file added
BIN
+139 KB
...xperiments/exp_9/tf_logs/events.out.tfevents.1704644122.Feriels-MacBook-Pro.local.56489.0
Binary file not shown.