Evolution Strategies in Optimization Problems

TL;DR

This paper demonstrates that simple evolution strategies are effective in solving complex optimal control problems by providing efficient approximations, bridging biological inspiration with practical optimization applications.

Contribution

It introduces the application of simple evolution strategies to optimal control, showing their effectiveness in obtaining good solutions for challenging problems.

Findings

Evolution strategies can efficiently approximate solutions in optimal control.

Simple strategies are effective for complex, recent optimal control problems.

The approach offers a practical alternative to traditional methods.

Abstract

Evolution Strategies are inspired in biology and part of a larger research field known as Evolutionary Algorithms. Those strategies perform a random search in the space of admissible functions, aiming to optimize some given objective function. We show that simple evolution strategies are a useful tool in optimal control, permitting to obtain, in an efficient way, good approximations to the solutions of some recent and challenging optimal control problems.