Stochastic Modeling of the Last Mile Problem for Delivery Fleet Planning

This paper presents a stochastic representation of the last mile problem that quantifies expected maintenance, regular labor, overtime labor, fuel, and carbon emission costs resulting from different delivery fleet options. The last mile delivery fleet planning model presented herein can be used in a decision framework to evaluate alternative delivery strategies involving fleet size and delivery frequency with information regarding cost, carbon emissions, service levels for available delivery hours, and payload capacity, as well as the transportation capacity needed to meet customer demand and lends itself well to performing what-if analyses.