New Model for Multi-Objective Evolutionary Algorithms

TL;DR

This paper introduces a new unified model for Multi-Objective Evolutionary Algorithms (MOEAs) that systematically encompasses various existing algorithms and facilitates the development of new ones.

Contribution

The paper proposes a novel unified model for MOEAs based on two sub-models, enabling systematic construction and analysis of MOEAs.

Findings

Decomposition of existing algorithms within the new model

Discussion of key issues in MOEA design

Potential for systematic development of new MOEAs

Abstract

Multi-Objective Evolutionary Algorithms (MOEAs) have been proved efficient to deal with Multi-objective Optimization Problems (MOPs). Until now tens of MOEAs have been proposed. The unified mode would provide a more systematic approach to build new MOEAs. Here a new model is proposed which includes two sub-models based on two classes of different schemas of MOEAs. According to the new model, some representatives algorithms are decomposed and some interesting issues are discussed.