1. ¹Royal Military College of Canada, Kingston, ON, Canada
2. ²GERAD and HEC, Montréal, Canada
3. ³CRT and HEC, Montréal QC, Canada
4. ⁴University of Birmingham, Birmingham, UK
5. ⁵Mathematical Institute, SANU, Belgrade, Serbia

Correspondence: G Laporte, CRT and HEC, 3000, Chemin de la Côte-Sainte-Catherine, Montréal, Quebec, Canada H3T 2A7; N Mladenovi, School of Mathematics, University of Birmingham, Birmingham, UK. E-mail: gilbert.laporte@hec.ca

Received January 2005; Accepted August 2006; Published online 3 January 2007.

Abstract

This paper examines the plant location problem under the objective of maximizing return-on-investment. However, in place of the standard assumption that all demands must be satisfied, we impose a minimum acceptable level on market share. The model presented takes the form of a linear fractional mixed integer program. Based on properties of the model, a local search procedure is developed to solve the problem heuristically. Variable neighbourhood search and tabu search heuristics are also developed and tested. Thus, a useful extension of the simple plant location problem is examined, and heuristics are developed for the first time to solve realistic instances of this problem.