A hybrid ant-tabu algorithm for solving a multistate flow network reliability maximization problem

Yi-Kuei Lin*, Cheng Ta Yeh, Pei Sheng Huang

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

11 Scopus citations

Abstract

Network reliability optimization for multistate flow networks (MFN) is an important issue for many system supervisors. Network reliability maximization for an MFN by determining the optimal component assignment, where a set of multistate components are ready to be assigned to the network, is a common problem. Previous research solved this problem by developing and applying genetic algorithm. Ant colony optimization (ACO) finds a good solution quickly by utilizing the experience of the proceeding ant but sometimes falls into local optimum. Tabu search (TS) adopts a tabu list to avoid searching in the same direction, and thus it explores other possible solutions. This strategy enlarges the search space. Therefore, we propose a hybrid ant-tabu (HAT) algorithm integrating the advantages of ACO and TS to solve this problem, where network reliability is evaluated in terms of minimal paths (MPs) and Recursive Sum of Disjoint Products. Experimental (RSDP) results show that the proposed HAT has better computational efficiency than several soft computing algorithms for networks with more than six MPs or 10 arcs.

Original languageEnglish
Pages (from-to)3529-3543
Number of pages15
JournalApplied Soft Computing Journal
Volume13
Issue number8
DOIs
StatePublished - 31 May 2013

Keywords

  • Component assignment
  • Hybrid ant-tabu
  • Multistate flow network
  • Network reliability maximization
  • Recursive Sum of Disjoint Products

Fingerprint Dive into the research topics of 'A hybrid ant-tabu algorithm for solving a multistate flow network reliability maximization problem'. Together they form a unique fingerprint.

Cite this