Freight Gateway Consolidation for Purolator International Using Integer Programming

This research reviews the northbound transportation network of a major Canadian courier, aiming to streamline U.S. transshipment operations. An integer programming model was developed to optimize gateway locations in the U.S. for cost-effective customer assignment and optimal truck routing to Canada. Four actionable recommendations from the model were implemented, resulting in annual cost savings surpassing USD 1.2 million, significantly exceeding initial estimates. In this research, our objective is to review the northbound transportation network from the United States to Canada for a major Canadian courier company and assess the efficiency of intermediate gateway locations. The aim is to streamline transshipment operations in the United States and reduce lead times on the Canadian side of the supply chain. We develop an integer programming formulation that takes into account hundreds of possible gateway locations in U.S. cities and demand from thousands of U.S. customers shipping to Canada as options for gateway facilities to perform the following: (i) ensure that gateways are assigned in a cost-effective manner and possibly consolidate gateways based on the customer list and shipping volumes, (ii) assign gateways to customers based on their respective destinations, and (iii) route an optimal number of trucks from gateways to major destinations in Canada. Our formulation led to four main reconfigurations that initiated a series of comprehensive analyses of the business processes at the company, which, in turn, resulted in four actionable recommendations. These recommendations were assessed by the company’s key decision makers with respect to their operational costs and other considerations and were estimated to yield a cost savings of around US$782,767 per year. According to our industrial partner, the realized cost savings after implementation surpassed the estimated amount by nearly US$500,000. History: This paper was refereed. Funding: E. Hassini was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC).