An analysis of a rail car unloading area for a consumer products manufacturer

Abstract

This paper reports on the results of an effort to design and analyse the rail car unloading area of Procter & Gamble's principal laundry detergent (soap powder) plant. In the first part of the study the design team established daily requirements for the number of raw material rail cars unloaded per day. The related combinatorial optimisation problem of assigning rail cars to positions on the platform and unloading equipment to rail cars was modelled as a mixed-integer nonlinear program. The inability of two standard commercial codes to find optimal solutions led to the development of a greedy randomised adaptive search procedure (GRASP). Accounting for the operational and physical limitations of the system, GRASP was used to determine the maximum performance that could be achieved under normal conditions. In the second part of the study alternative designs were proposed for meeting an expected 14% increase in demand over the next few years. The analytic hierarchy process in conjunction with a standard scoring model was used to rank the evaluation criteria and to select the preferred alternative. A worst-case analysis of the top candidate confirmed its performance capabilities.