CEC2020 Algorithms

Environment to comparision of evolutionary algorithms based on CEC benchmarks.

Constaints Python implementations of the best algorithms from CEC2020:

CEC2020 - Special Session & Competitions on Real-World Constrained Optimization

• COLSHADE
• EnMODE
• esCMAgES

CEC2020 - Competition on Single Objective Bound Constrained Numerical Optimization

• AGSK
• IMODE
• j2020

Run

There are some operations you can do in this environment:

Optimization

To run optimization of functions type in terminal:

python optimize.py BENCHMARK ALGORITH FUNCTION DIMENSIONALITY RUNS [-f BUDGET] [-p PATH]

Aggregate results

To aggreagate results of optimization and calculate statistics type in terminal:

python calc_stats.py BENCHMARK ALGORITH [-f [FUNCTIONS ...]] [-l] [-d DIM] [-p PATH] [-s] [-v]

ECDF curves

To draw ECDF curves based on results of optimization and calculate rankings type in terminal:

python ecdf.py BENCHMARK -a [ALGORITHMS ...] [-f [FUNCTIONS ...]] [-l] [-d DIM] [-c] [-p PATH] [-b BUDGET] [-s] [-v]

Wilocoxon test

To perform Wilcoxon test between two algorithms type in terminal:

python wilcoxon.py ALGORITHM1 ALGORITHM2 [-f [FUNCTIONS ...]] [-l] [-d DIM] [-p PATH] [-s] [-v]

CEC2020 rankings

To calculate CEC2020 rankings type in terminal:

python calc_rankings.py BENCHMARK -a [ALGORITHMS ...] [-s] [-v]