Approximate Solving of Large p-Median Problems

This paper deals with an approximate approach to the p-median problem. Large instances of the p-median problem can be often met when a real public service system is designed and the discomfort of its customers is minimized. If a large instance of the p-median problem is described by a location-allocation model, then the size of a model often exceeds the acceptable limit for available optimization software. Using presented approximate approach, a user can avoid this obstacle paying for it by a loss of accuracy. In this paper, we present both exact and approximate approaches to the p-median problem and try to find how the loss of accuracy depends on parameters of the approximate approach. Keywords— Public Service System Design, Large pMedian Problem, Covering Model, Approximate Approach