The aim of this project is to implement both serial and parallel version of an Evolutionary Algorithm for Large-Scale Black Box Optimization. The parallelization is done using OpenMP.
In this project, the proposed evolutionary strategy algorithm using sparse plus low rank model for large-scale black box optimization was implemented and was parallelized using the OpenMP API. The proposed algorithm consists of two methods - (i) Rank-One evolution strategy(R1-ES) which uses a single principle direction for optimization and its extended version (ii) Rank-m evolution strategy(Rm-ES) which uses multiple principle directions. Both of these algorithms were parallelized and their convergence and speedup were analysed w.r.t some critical parameters.
GCC/G++ Compiler(https://gcc.gnu.org)OpenMP(https://www.openmp.org/resources/openmp-compilers-tools)
Contributions in shape of [Pull Requests] are always welcome.