Minisum location problem with farthest Euclidean distances

The paper formulates an extended model of Weber problem in which the customers are represented by convex demand regions. The objective is to generate a site in R2 that minimizes the sum of weighted Euclidean distances between the new facility and the farthest points of demand regions. This location problem is decomposed into a polynomial number of subproblems: constrained Weber problems (CWPs). A projection contraction method is suggested to solve these CWPs. An algorithm and the complexity analysis are presented. Three techniques of bound test, greedy choice and choice of starting point are adopted to reduce the computational time. The restricted case of the facility is also considered. Preliminary computational results are reported, which shows that with the above three techniques adopted the algorithm is efficient.