Optimizing Epochal Evolutionary Search: Population-Size Independent Theory

TL;DR

This paper develops a mathematical framework to optimize evolutionary search by analyzing epochal dynamics, providing estimates and bounds on parameters to minimize the number of evaluations needed to find the global optimum.

Contribution

It introduces a population-size independent theoretical model for epochal evolution, enabling optimized parameter selection for efficient search.

Findings

Derived estimates for the number of evaluations to reach the global optimum.

Provided bounds on evolutionary parameters to minimize search effort.

Analyzed epochal dynamics in natural and artificial evolution.

Abstract

Epochal dynamics, in which long periods of stasis in population fitness are punctuated by sudden innovations, is a common behavior in both natural and artificial evolutionary processes. We use a recent quantitative mathematical analysis of epochal evolution to estimate, as a function of population size and mutation rate, the average number of fitness function evaluations to reach the global optimum. This is then used to derive estimates of and bounds on evolutionary parameters that minimize search effort.