A Time-Constrained Vehicle Routing Problem with a Heterogeneous Fleet: Algorithms and Analysis

In this paper, we consider a new variant of the Time-Constrained Heterogeneous Vehicle Routing Problem (TCHVRP). In this problem, the cost and travel time of any given arc vary by vehicle type within a heterogeneous eet. In particular, we make no assumptions about Pareto dominance across vehicles; nor do we assume that cost and time are correlated. Our research is motivated by situations in which existing eets are evolving to incorporate hybrid vehicles in addition to their existing vehicle types; for many vehicles, the cost per mile depends heavily on the type of driving (such as highway versus city). We formulate TCHVRP as a path-based model, which we solve using column generation. We introduce several di erent methods to solve the pricing problem. We conclude by conducting empirical analyses to assess both the performance of the proposed approach.