Multi-criteria mathematical model for designing the distribution network of a consumer goods company

This paper describes a multiple criteria mixed-integer linear program used for designing the best possible supply chain distribution network for a consumer goods company. The model determines the optimal configuration of the manufacturing plants, distributors and customers in the distribution network. The model is intended for making tactical decisions for designing distribution networks, or more specifically, for designing the flow of products from the manufacturing plants to the customers. The customers have four options for receiving products in this model. Products can be supplied from (1) the regional distribution center (DC), (2) the manufacturing plant, (3) an independent distributor who is supplied from the regional DC, or (4) an independent distributor who is supplied directly from a manufacturing plant. The model selects the best option for each customer/distributor based on several criteria: profit, lead time, power, credit performance, and distributors' reputation. The model is validated with real data from a consumer goods company to show its functionality. To account for variability in demand, the model is run under multiple scenarios and the results are analyzed to obtain the best solution. The company uses two DC's located in the same region, but the model assumes only one regional DC with infinite capacity. With the proposed network, we show that one of the regional DC's could be eliminated and distribution costs would be reduced from 12% to 3% of the Net Sales (approximately a monthly reduction of $574,000 in distribution expenses).