Skip to main navigation Skip to search Skip to main content

Multi-objective Techniques for Single-Objective Local Search: A Case Study on Traveling Salesman Problem

Research output: Chapters, Conference Papers, Creative and Literary WorksRGC 32 - Refereed conference paper (with host publication)peer-review

Abstract

In this paper, we show that the techniques widely used in multi-objective optimization can help a single-objective local search procedure escape from local optima and find better solutions. The Traveling Salesman Problem (TSP) is selected as a case study. Firstly the original TSP ƒ0 is decomposed into two TSPs ƒ1 and ƒ2 such that ƒ0ƒ1ƒ2 . Then we propose the Non-Dominance Search (NDS) method which applies the non-domination concept on (ƒ1ƒ2 to guide a local search out of the local optima of ƒ0 . In the experimental study, NDS is combined with Iterated Local Search (ILS), a well-known metaheuristic for the TSP. Experimental results on some selected TSPLIB instances show that the proposed NDS can significantly improve the performance of ILS.
Original languageEnglish
Title of host publicationEvolutionary Multi-Criterion Optimization
Subtitle of host publication10th International Conference, EMO 2019, Proceedings
EditorsKalyanmoy Deb, Erik Goodman, Carlos A. Coello Coello, Kathrin Klamroth, Kaisa Miettinen, Sanaz Mostaghim, Patrick Reed
PublisherSpringer Nature Switzerland AG
Pages114-125
ISBN (Electronic)9783030125981
ISBN (Print)9783030125974
DOIs
Publication statusPublished - Mar 2019
Event10th International Conference on Evolutionary Multi-Criterion Optimization, EMO 2019 - East Lansing, United States
Duration: 10 Mar 201913 Mar 2019

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PublisherSpringer
VolumeLNCS 11411
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference10th International Conference on Evolutionary Multi-Criterion Optimization, EMO 2019
PlaceUnited States
CityEast Lansing
Period10/03/1913/03/19

Research Keywords

  • Multi-objective optimization
  • Traveling Salesman Problem
  • Local search
  • Metaheuristic
  • Iterated Local Search

Fingerprint

Dive into the research topics of 'Multi-objective Techniques for Single-Objective Local Search: A Case Study on Traveling Salesman Problem'. Together they form a unique fingerprint.

Cite this