
Accession Number : AD0833993
Title : EXACT SOLUTION OF THE SITE SELECTION PROBLEM BY MIXED INTEGER PROGRAMMING.
Descriptive Note : Technical rept.,
Corporate Author : STANFORD RESEARCH INST MENLO PARK CA
Personal Author(s) : Gray, Paul
Report Date : NOV 1967
Pagination or Media Count : 136
Abstract : Algorithms for solving siteselection and similar fixed charge problems with upper bound constraints are presented. The basic approach is to formulate the problem as a mixed integer program and to solve these programs by decomposing them into a master integer program and a series of subproblems which are linear programs. To reduce the number of vertices to be examined to manageable proportions, the boundandscan algorithm by F. S. Hillier was adapted to the fixed charge problem. Algorithms are presented for four classes of problems: (1) Fixed charge problem with linear variable costs and a fixed charge for each variable at nonzero level. (2) Problem 1 with separable concave or convex variable costs. (3) A warehouse location problem in which variable costs and constraints are of the transportation type. A fixed charge is associated with each warehouse opened. (4) The fixed charge transportation problem in which a fixed charge is associated with each route rather than with each warehouse. Computational results for Problems 1, 2 and 4 are presented. (Author)
Descriptors : *MATHEMATICAL PROGRAMMING), (*SITE SELECTION, (*PROBLEM SOLVING, SITE SELECTION), ALGORITHMS, OPTIMIZATION, MATRICES(MATHEMATICS), NONLINEAR SYSTEMS, COSTS, TRANSPORTATION, THEOREMS, WAREHOUSES, LINEAR PROGRAMMING.
Subject Categories : Theoretical Mathematics
Operations Research
Distribution Statement : APPROVED FOR PUBLIC RELEASE