Sequential Zone Adjustment for Approximate Solving of Large p-Median Problems

The p-median problems are used very often for designing optimal structure of most public service systems. Since particular models are characterized by big number of possible facility locations, current exact approaches often fail. This paper deals with an approximate approach based on specific model reformulation. It uses an approximation of the network distance between a service center and a customer by some of pre-determined distances. The pre-determined distances are given by dividing points separating the range of possible distances. Deployment of these points influences the accuracy of the approximation. To improve this approach, we have developed a sequential method of the dividing point deployment.We study here the effectiveness of this method and compare the results to the former static method.