A novel mutation operator based on the union of fitness and design spaces information for Differential Evolution

TL;DR

This paper introduces a new mutation operator for Differential Evolution that combines information from both fitness and design spaces, significantly enhancing optimization performance on benchmark problems.

Contribution

A universal differential evolution method is proposed that intelligently combines design and fitness space information for mutation vector selection.

Findings

UDE outperforms methods using only one criterion on CEC2005 benchmarks.

Significant performance improvements over traditional DE methods.

Experimental results validate the effectiveness of combining design and fitness spaces.

Abstract

Differential Evolution (DE) is one of the most successful and powerful evolutionary algorithms for global optimization problem. The most important operator in this algorithm is mutation operator which parents are selected randomly to participate in it. Recently, numerous papers are tried to make this operator more intelligent by selection of parents for mutation intelligently. The intelligent selection for mutation vectors is performed by applying design space (also known as decision space) criterion or fitness space criterion, however, in both cases, half of valuable information of the problem space is disregarded. In this article, a Universal Differential Evolution (UDE) is proposed which takes advantage of both design and fitness spaces criteria for intelligent selection of mutation vectors. The experimental analysis on UDE are performed on CEC2005 benchmarks and the results stated that UDE significantly improved the performance of differential evolution in comparison with other methods that only use one criterion for intelligent selection.