Multilocation plant sizing and timing problem: Analysis of spatial decomposition procedure

The multilocation plant sizing and timing problem (MLPSTP) is an important strategic management problem which involves determination of size, location, time-phasing and utilization of projects to meet the demands of spatially distributed customers. The objective for the analysis of such a problem has been to minimize the net present worth of the capital costs as well as the streams of operating and transportation costs. MLPSTP is a mixed integer programming problem involving both binary and continuous variables. The continuous variables relate to the transportation of commodities from site to market. The binary variables problem relates to the size, location and time-phasing of the projects. In this paper, an attempt has been made to analyse the problem by using hierarchical decomposition procedures. The approach has been to construct a two-level hierarchy; the first level identifying the distinct geographical areas as independent subproblems, and the second level being the co-ordination level which aims at synthesizing the solutions of the area subproblems.