Vehicle Routing Optimization

Transportation costs, including all stages of manufacturing and distribution, account for 10% to 20% of the final cost of a consumer good. Obviously, methods that would reduce this cost are highly desirable for any organization. The problems faced, in planning and carrying out the transportation of goods, are very complex, and therefore lend themselves well to optimization methods. Optimization methods have been successfully applied to many of the logistics problems faced by the trucking industry, and in general shipping. This paper describes an optimization model of a classic routing problem in the trucking industry: the “Vehicle Routing Problem with Time Windows” (VRPTW). The basic vehicle routing problem was first described by Danzig and Ramser over 40 years ago, but implementation of decision support tools based on optimization models have only appeared in widespread use over the past 15 years. The basic objective of a vehicle routing problem is to select the best set of routes so that overall cost is minimized. Data for the model presented in this report was provided by Con-Way Transportation Services, Inc., specifically from their local city route deliveries. The total city route problem of Con-Way was modeled in two parts: Trip Optimization and Route Optimization. In the first part, the total set of shipments was partitioned into sub-sets of individual truckloads. The model took into account truck capacity, as well as total miles traveled per truckload. The objective being to minimize the number of trucks required. In the second part, the best route for each truckload was determined. Time windows for each shipment were taken into account, and the objective was to minimize total miles traveled