Inheritance-Based Diversity Measures for Explicit Convergence Control in Evolutionary Algorithms

TL;DR

This paper explores explicit, quantifiable diversity measures in evolutionary algorithms that do not rely heavily on domain knowledge, introducing genealogical diversity to improve global optimization.

Contribution

It introduces and analyzes inheritance-based diversity measures, including genealogical diversity, for better convergence control without extensive domain knowledge.

Findings

Diversity measures can prevent premature convergence.

Genealogical diversity enhances global search capabilities.

Explicit diversity control improves optimization outcomes.

Abstract

Diversity is an important factor in evolutionary algorithms to prevent premature convergence towards a single local optimum. In order to maintain diversity throughout the process of evolution, various means exist in literature. We analyze approaches to diversity that (a) have an explicit and quantifiable influence on fitness at the individual level and (b) require no (or very little) additional domain knowledge such as domain-specific distance functions. We also introduce the concept of genealogical diversity in a broader study. We show that employing these approaches can help evolutionary algorithms for global optimization in many cases.