A New Mechanism for Maintaining Diversity of Pareto Archive in Multiobjective Optimization

TL;DR

This paper presents a novel mechanism for maintaining diversity in Pareto archives within multiobjective optimization, demonstrating improved distribution without sacrificing convergence compared to existing algorithms.

Contribution

A new diversity-preserving mechanism integrated with a micro-genetic algorithm, outperforming NSGA-II, SPEA2, and IBEA in distribution quality.

Findings

Achieves uniform Pareto set distribution

Maintains convergence quality

Effective for small populations

Abstract

The article introduces a new mechanism for selecting individuals to a Pareto archive. It was combined with a micro-genetic algorithm and tested on several problems. The ability of this approach to produce individuals uniformly distributed along the Pareto set without negative impact on convergence is demonstrated on presented results. The new concept was confronted with NSGA-II, SPEA2, and IBEA algorithms from the PISA package. Another studied effect is the size of population versus number of generations for small populations.