A hierarchical evolutionary approach to multi-objective optimization

Abstract

This paper describes a hierarchical evolutionary \napproach to Pareto-based multi-objective optimization. Using the SEAMO algorithm (a simple evolutionary algorithm for multiobjective optimization) as a basis, it demonstrates how it is possible to obtain a better spread of results if subpopulations of various sizes are used in a simple hierarchical framework. Three alternative hierarchical models are tried and the results compared.