Hybridizing the Pareto Multi-Objective Optimization Evolutionary Algorithms by Means of Multi-Objective Local Search

Author

Haidine, Abdelfatteh ; Lehnert, Ralf

Author_Institution

Tech. Univ. Dresden, Dresden

fYear

2008

fDate

10-12 Sept. 2008

Firstpage

314

Lastpage

319

Abstract

Hybridizing of evolutionary algorithms (EA) by means of local search has shown considerable performance improvement in single-objective optimization (SOO) field. The fine search in the neighborhood of the EA individuals (solutions) allows a fine exploration of the solution space. This paper investigates the application and the evaluation of the hybridizing mechanism of the EAs in the multi-objective optimization (MOO) domain. For this hybridizing, two types of multi-objective local search (MOLS) are used; namely large- and narrow-MOLS. Among numerous possible multi-objective optimization EAs (MOEAs), the Pareto-based variants have been considered. The performance of hybrid MOO variants are evaluated by solving the multi-objective knapsack problem.

Keywords

Pareto optimisation; evolutionary computation; knapsack problems; search problems; Pareto multiobjective optimization; evolutionary algorithm; knapsack problem; multiobjective local search; Approximation algorithms; Constraint optimization; Convergence; Evolutionary computation; Genetic algorithms; Hybrid intelligent systems; Hybrid power systems; Pareto optimization; Sorting; Wheels; evolutionary algorithm; hybrid algorithm; multi-objective algorithm; multi-objective local search;

fLanguage

English

Publisher

ieee

Conference_Titel

Hybrid Intelligent Systems, 2008. HIS '08. Eighth International Conference on

Conference_Location

Barcelona

Print_ISBN

978-0-7695-3326-1

Electronic_ISBN

978-0-7695-3326-1

Type

conf

DOI

10.1109/HIS.2008.104

Filename

4626648