Skip to content

Latest commit

 

History

History
16 lines (15 loc) · 923 Bytes

README.md

File metadata and controls

16 lines (15 loc) · 923 Bytes

Evolution Strategies for Constraint Optimization Problems

The main branch of this repository contains all sources for experimentation with the Matrix Adaptation Evolution Strategy in constrained environments.

As a first step you need a suitable Python environment.
Anaconda users may create an admissible environment provided by the yml-file 'es4cop_env'.
Just execute
conda env create -f es4cop_env.yml
and
conda activate es4cop
(Note: Basically, only Pyhton3.7+, Numpy, Matplotlib.Pyplot and Seaborn are required.)

The resources for the experiments are provided in the py-files maes.py and text-functions.py.

  • text-functions.py contains the COP formulations
  • maes.py contains the two variants of the MA-ES and various constraint handling routines

Usage is demonstrated in the Jupyter notebook templates `MAES-template.ipynb` and `eMAES-template.ipynb`.