Exact and Approximate Solutions to the Multisource Weber Problem

Abstract

The concern of modern spatial economists with the optimal location of variable points (sources) in 2-space with respect to a set of fixed points (sinks), when the co-ordinates of the sources may vary continuously, dates from the publication of Alfred Weber’s work in industrial location theory [15]. Weber’s analysis was largely confined to the location of a single source, and, although he offered no method of solution, recent work has led to an efficient algorithm for exact solution of the problem (the single-source algorithm). The present paper treats the multisource Weber problem and presents (1) a branch-and-bound algorithm for the exact solution of the problem, which, to the best of our knowledge, is original (MULTIWEB), and (2) an approximate algorithm, to be used in support of MULTIWEB or in its place when appropriate, for which we claim no priority (CROSSCUT).

Co-authored by Richard M. Soland and published originally in Mathematical Programming, 3 (1972), pp. 193–209, and reproduced with permission of Elsevier Science Publishers.For a more detailed discussion of the analyses contained in the present paper, the computational results, and for detailed program listings, the reader is referred to [8]. Tapes for the two programs discussed in the present paper are available at cost of reproduction from Institute for Defense Analyses, Alexandria, Virginia, USA.