Cited 4 time in webofscience Cited 6 time in scopus

A hybrid multiobjective evolutionary algorithm: Striking a balance with local search

Title
A hybrid multiobjective evolutionary algorithm: Striking a balance with local search
Authors
Ahn, CW[Ahn, Chang Wook]Kim, E[Kim, Eungyeong]Kim, HT[Kim, Hyun-Tae]Lim, DH[Lim, Dong-Hyun]An, J[An, Jinung]
DGIST Authors
An, J[An, Jinung]
Issue Date
2010-12
Citation
Mathematical and Computer Modelling, 52(11-12), 2048-2059
Type
Article
Article Type
Article; Proceedings Paper
Keywords
Computational CostsDiversity PerformanceEvolutionary AlgorithmsEvolutionary Multiobjective OptimizationFitness ValuesHealthInteger ProgrammingKnapsack ProblemKnapsack ProblemsKnowledge Based SystemsLocal SearchMulti-Objective OptimizationMulti-Objective Optimization ProblemMulti ObjectiveMulti Objective Evolutionary AlgorithmsNon-Dominated SolutionsPareto FrontSearch SpacesSearch TimeWeighted Fitness
ISSN
0895-7177
Abstract
This paper presents a hybrid multiobjective evolutionary algorithm (HMEA) that efficiently deals with multiobjective optimization problems (MOPs). The aim is to discover new nondominated solutions in the neighborhood of the most promising individuals in order to effectively push individuals toward the global Pareto front. It can be achieved by bringing the strength of an adaptive local search (ALS) to bear upon the evolutionary multiobjective optimization. The ALS is devised by combining a weighted fitness strategy and a knowledge-based local search which does not incur any significant computational cost. To be more exact, the highly converged and less crowded solutions selected in accordance with the weighted fitness values are improved by the local search, thereby helping multiobjective evolutionary algorithms (MEAs) to economize on the search time and traverse the search space. Thus, the proposed HMEA that transplants the ALS to the framework of MEAs can achieve higher proximity and better diversity of nondominated solutions. To show the utility of HMEA, the ALS for multiobjective knapsack problems (MKPs) is developed by exploiting the problem's knowledge. Experimental results on the MKPs have provided evidence for its effectiveness as regards the proximity and the diversity performances. © 2010 Elsevier Ltd.
URI
http://hdl.handle.net/20.500.11750/1627
DOI
10.1016/j.mcm.2010.06.007
Publisher
PERGAMON-ELSEVIER SCIENCE LTD
Related Researcher
Files:
There are no files associated with this item.
Collection:
Convergence Research Center for Wellness1. Journal Articles


qrcode mendeley

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

BROWSE