USE OF CONTINUOUS SPACE MODELING TO ESTIMATE FREIGHT DISTRIBUTION COSTS

Abstract This paper develops an analytical framework for estimating the cost of distributing freight from one origin to many destinations. The origin and destinations each represent specific geographic locations, e.g., a supplier shipping to different retail outlets. Analytical estimates of freight distribution costs in the literature assume destinations have equal demand and are randomly located with uniform density. This research generalizes the techniques to account for spatial variations in the demand and density of destinations. General equations are derived for the two main cost components: the number of destination stops per load, and the distance traveled per load. Numerical examples are considered and compared to the simplest case of equal demand and uniform density. Results suggest that, for many practical situations, equations for the simplest case provide sufficiently accurate cost estimates. The analysis uses a continuous space modeling approach, which requires only the spatial density of destinations and the average and variance of demand, rather than data on the location and demand of each destination individually. This approach allows distribution costs to be determined analytically in terms of a few easily measured parameters.