A parallel implementation of the covariance matrix adaptation evolution strategy

TL;DR

This paper presents a parallel implementation of the CMA-ES algorithm to improve its speed for derivative-free optimization tasks, demonstrating enhanced performance on benchmark problems.

Contribution

The paper introduces a parallel version of CMA-ES and analyzes its performance, addressing the computational complexity for faster optimization.

Findings

Parallel CMA-ES reduces execution time

Improved performance on benchmark problems

Effective for derivative-free optimization

Abstract

In many practical optimization problems, the derivatives of the functions to be optimized are unavailable or unreliable. Such optimization problems are solved using derivative-free optimization techniques. One of the state-of-the-art techniques for derivative-free optimization is the covariance matrix adaptation evolution strategy (CMA-ES) algorithm. However, the complexity of CMA-ES algorithm makes it undesirable for tasks where fast optimization is needed. To reduce the execution time of CMA-ES, a parallel implementation is proposed, and its performance is analyzed using the benchmark problems in PythOPT optimization environment.