Authors: Dun-wei Gong, Yong Zhang
Addresses: School of Information and Electronic Engineering, China University of Mining and Technology, Xunzhou 221008, China. ' School of Information and Electronic Engineering, China University of Mining and Technology, Xunzhou 221008, China
Abstract: Multi-project location problems, a kind of combinational optimisation problem, widely exist in real-world applications. However, the existing methods can only deal with the case of exact objective profits. Aiming at solving the case of interval objective profits, we present a novel method based on an improved particle swarm optimisation in this paper. We firstly convert a multi-project location problem with interval objective profits into one with exact parameters by designing an appropriate penalty function. Based on the equivalent probability matrix, we then develop an improved particle swarm optimisation with reverse mutation, suitable for discrete variables, to optimise the converted problem, where the role of the reverse mutations is to escape these particles from local optima of the problem. We finally validate the feasibility and efficiency of the proposed method by optimising a location problem with 15-project and 13-site.
Keywords: particle swarm optimisation; PSO; multi-project location; interval profit; reverse mutation; combinational optimisation.
International Journal of Modelling, Identification and Control, 2009 Vol.8 No.4, pp.335 - 343
Available online: 09 Dec 2009 *Full-text access for editors Access for subscribers Purchase this article Comment on this article