Combining two-phase local search with multi-objective ant colony optimization

Chun Wa Leung, Sin Chun Ng, Andrew K. Lui

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Multi-objective Ant Colony Optimization (MOACO) is a popular algorithm in solving the multi-objective combinational optimization problem. Many variants were introduced to solve different types of multi-objective optimization problem. However, MOACO does not guarantee to generate a good approximation of solution in a predefined termination time. In this paper, Two-Phase Local Search (TPLS) was introduced to cooperate with the MOACO, a two-phase strategy that creates a very good approximation of Pareto Front at the beginning of the algorithm and then further explores the Pareto Front iteratively. We also propose Iterated Local Search – Variable Neighborhood Search (ILS-VNS) as the first phase in TPLS, an iterative improvement process that allows finding improving solutions from adaptively sized neighborhood space. A series of experiments were performed to investigate the performance improvement on the solutions. At the same time, we studied the effect of Weighted Local Search (WLS) and Pareto Local Search (PLS) in the proposed algorithm. The results showed that the newly proposed algorithm obtains a larger area on hypervolume space and exhibits a significantly larger accuracy rate in obtaining the true Pareto-optimal solutions. Additionally, a statistical testing was also performed to verify the significance of the result.

Original languageEnglish
Title of host publicationNeural Information Processing - 25th International Conference, ICONIP 2018, Proceedings
EditorsAndrew Chi Sing Leung, Seiichi Ozawa, Long Cheng
Pages564-576
Number of pages13
DOIs
Publication statusPublished - 2018
Event25th International Conference on Neural Information Processing, ICONIP 2018 - Siem Reap, Cambodia
Duration: 13 Dec 201816 Dec 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11302 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference25th International Conference on Neural Information Processing, ICONIP 2018
Country/TerritoryCambodia
CitySiem Reap
Period13/12/1816/12/18

Keywords

  • Bi-objective traveling salesman problem
  • Hybridization
  • Iterated local search – Variable neighborhood search
  • Multi-objective ant colony optimization
  • Two-phase local search

Fingerprint

Dive into the research topics of 'Combining two-phase local search with multi-objective ant colony optimization'. Together they form a unique fingerprint.

Cite this