A novel Bi-Ant colony optimization algorithm for solving multi-objective service selection problem

Issue title: ICNC-FSKD 2015

Guest editors: Zheng Xiao and Kenli Li

Article type: Research Article

Authors: Huang, Liping^{a; b; *} | Zhang, Bin^b | Yuan, Xun^c | Zhang, Changsheng^b | Ma, Anxiang^b

Affiliations: [a] Software College, Northeastern University, Shen Yang, China | [b] College of Information Science & Engineering, Northeastern University, Shen Yang, China | [c] Institute of Information Science & Engineering, Shenyang Ligong University, Shen Yang, China

Correspondence: [*] Corresponding atuthor. Liping Huang, Software College, Northeastern University, Shen Yang, China. Tel.: +86 24 83680517; Fax: +86 24 83680522; E-mail: huanglp@swc.neu.edu.cn.

Abstract: The multi-objective service selection problem is a basic problem in Service Computing and it is NP-Hard. This paper proposes a novel Bi-Ant colony optimization (NBACO) algorithm for this problem. Two objective functions related to response time and cost attributes are considered. For each objective, a heuristic function and a pheromone updating policy are defined against the characteristics of this problem. Then, a combined state transition rule is designed based on them. It uses preposition skyline query (PSQ) algorithm for each service class to reduce the candidate services at the beginning of NBACO. The algorithm has been tested in nine cases and compared to the related MOACO algorithm and Co-Evolution algorithm for this problem. The efficiency of NBACO is greatly improved by using PSQ. The result demonstrates that our approach is effective and better than MOACO and Co-Evolution.

Keywords: Multi-objective, service selection, PSQ, NBACO

DOI: 10.3233/JIFS-169018

Journal: Journal of Intelligent & Fuzzy Systems, vol. 31, no. 2, pp. 873-884, 2016