Heuristic cluster algorithm for multiple facility location-allocation problem

The multiple facility location-allocation problem consists of finding the optimal set of location points to establish the facility centers at them and the allocation of every demand point to a facility center. The problem can be solved by determining an optimal partition of the demand point set and solving the corresponding single facility location problems. The authors propose a general method based on Cluster Analysis for obtaining a heuristic partition and provide the specific algorithms for the standard models. They compare these procedures experimentally with other known heuristics on different sized randomly generated instances of the p-median problem. Its low memory requirements, its efficiency and the high degree of optimality attained mean that is a method which is particularly suited for using with personal computers.