Algorithm for Modular-Capacitated Multi-Period Plant Location Problem with capacity closure constraint

Abstract

Selection of location for manufacturing plants is a strategic decision for an organization. Shifts in customer demand during the plant’s lifespan can alter the attractiveness of a particular location, turning an optimal location of one period into a strategic blunder for the future. Closure or relocation of plants may be unviable, due to external factors and these inefficient locations would result in excess transportation costs, which cannot be offset, no matter how well the production plans or inventory are optimized in the operational level plans. The complexity of modeling such problems has limited much of the traditional facility location research to simplified static (single-period) models. This paper presents an algorithm to generate the optimal sequence for opening plants and installing modular capacity units across locations during a multi-period planning horizon. The objective is to achieve the lowest cumulative cost of transportation and capital investment. The algorithm was applied to a randomly generated set of locations (50 customers and 20 candidate plants) over a 10 year demand horizon. The multi-period model achieved a capacity sequence with a cumulative cost 3.2% lower than the year-on-year planned sequence. To demonstrate the algorithm on an industry application, it was applied for the Indian automobile industry. This industry is a good candidate for the model as it has high transportation costs and capital intensive capacity (built in modular assembly lines). Also, it has a large ecosystem of vendors, workmen and their families in the vicinity of the plants, which makes closure or relocation of capacity unviable.