Schemata-driven multi-objective optimization

Abstract

This paper investigates the exploitation of non-dominated sets' schemata in guiding multi-objective optimization. Schemata capture the similarities between solutions in the non-dominated set. They also reflect the knowledge acquired by multi-objective evolutionary algorithms. A schemata-driven genetic algorithm as well as a schemata-driven local search algorithm are described. An experimental study to evaluate the suggested approach is then conducted. © Springer-Verlag Berlin Heidelberg 2003.