본문 바로가기
  • Home

Integer Programming-based Local Search Techniques for the Multidimensional Knapsack Problem

  • Journal of The Korea Society of Computer and Information
  • Abbr : JKSCI
  • 2012, 17(6), pp.13-27
  • Publisher : The Korean Society Of Computer And Information
  • Research Area : Engineering > Computer Science

Junha Hwang ORD ID 1

1금오공과대학교

Accredited

ABSTRACT

Integer programming-based local search(IPbLS) is a kind of local search based on simple hill-climbing search and adopts integer programming for neighbor generation unlike general local search. According to an existing research [1], IPbLS is known as an effective method for the multidimensional knapsack problem(MKP) which has received wide attention in operations research and artificial intelligence area. However, the existing research has a shortcoming that it verified the superiority of IPbLS targeting only largest-scale problems among MKP test problems in the OR-Library. In this paper, I verify the superiority of IPbLS more objectively by applying it to other problems. In addition, unlike the existing IPbLS that combines simple hill-climbing search and integer programming, I propose methods combining other local search algorithms like hill-climbing search, tabu search, simulated annealing with integer programming. Through the experimental results, I confirmed that IPbLS shows comparable or better performance than the best known heuristic search also for mid or small-scale MKP test problems.

Citation status

* References for papers published after 2023 are currently being built.

This paper was written with support from the National Research Foundation of Korea.