Contributions of operations research to location analysis

Location problems which can be quantified as optimization problems are natural candidates for operations research approaches, and many such problems have been studied using mathematical programming methodology in the last 30 years or so. This paper attempts to give a largely non-technical overview of some of this activity. A number of actual problems are discussed, and then models of these problems are presented. Models are classified as planar, network, and mixed integer programming models, and methodology for solving such models is outlined. A particularly important location problem, known as the simple plant location problem, or warehouse location problem, is discussed in some detail, with the emphasis placed on solution approaches. The three model classes are compared in terms of four important attributes: realism, data requirements, computational requirements, and difficulty of explanation; no single class is best in all four attributes. Some opportunities for future work, particularly the need for generally available software, are identified. References are given for further reading, including texts and review papers.