EVOLUTIONARY MULTI-OBJECTIVE OPTIMIZATION VIA DIFFERENTIAL EVOLUTION

Abstract

Multi-objective optimization is extensively applied in many fields like engineering, logistics, economics, bioinformatics, finance or any other real-life applications that involves two or more conflicting objectives that need to be optimized simultaneously. Differential evolution is a simple but powerful evolutionary optimization algorithm of high popularity with many successful applications. The primary aim of this thesis is to develop novel differential evolution algorithms in the context of multi-objective optimization and to implement the algorithms to solve both theoretical and real-life application problems with vastly different characteristics and representation schemes. The optimization performance of the novel differential evolution algorithms is then studied for scalable, many-objective and permutation-based multi-objective optimization problems.