Agent-Based Co-Evolutionary Techniques for Solving Multi-Objective Optimization Problems

Book Description

During last 25 years multi-objective optimization has been in the limelight of researchers. Because of practical importance and applications of multi-objective optimization as the most natural way of decision making and real-life optimizing method--growing interests of researchers in this very field of science was a natural consequence and extension of previous research on single-objective optimization techniques. Unfortunately, when searching for the approximation of the Pareto frontier, classical computational methods often prove ineffective for many (real) decision problems. The corresponding models are too complex or the formulas applied too complicated, or it can even occur that some formulations must be rejected in the face of numerical instability of available solvers. Also, because of such a specificity of multi-objective optimization (especially when--as in our case--we are considering multi-objective optimization in the Pareto sense) that we are looking for the whole set of nondominated solutions rather than one single solution--the special attention has been paid on population-based optimization techniques and if so, the most important techniques turned out here to be evolutionary-based methods. Research on applying evolutionary-based methods for solving multi-objective optimization tasks resulted in developing a completely new (and now commonly and very well known) science field: evolutionary multi-objective optimization (EMOO). To confirm above sentences, it is enough to mention statistics regarding at least the number of conference and journal articles, PhD thesis, conferences, books etc. devoted to EMOO and available at http://delta.cs.cinvestav.mx/~coello/EMOO. After the first stage of research on EMOO when plenty of algorithms were proposed1, simultaneously with introducing in early 2000s two the most important EMOO algorithms