Search

Although increasing amounts of transaction data make it possible to characterize uncertainties surrounding customer service requests, few methods integrate predictive tools with prescriptive optimization procedures to meet growing demand for small-volume urban transport services. We incorporate temporal and spatial anticipation of service requests into approximate dynamic programming (ADP) procedures to yield dynamic routing policies for the single-vehicle routing problem with stochastic service requests, an important problem in city-based logistics. We contribute to the routing literature as well as to the field of ADP. We combine offline value function approximation (VFA) with online rollout algorithms resulting in a high-quality, computationally tractable policy. Our offline–online policy enhances the anticipation of the VFA policy, yielding spatial and temporal anticipation of requests and routing developments. Our combination of VFA with rollout algorithms demonstrates the potential benefit of using offline and online methods in tandem as a hybrid ADP procedure, making possible higher-quality policies with reduced computational requirements for real-time decision making. Finally, we identify a policy improvement guarantee applicable to VFA-based rollout algorithms, showing that base policies composed of deterministic decision rules lead to rollout policies with performance at least as strong as that of their base policy.

Many real-world vehicle routing problems are dynamic optimization problems, with customer requests arriving over time, requiring a repeated reoptimization. In this paper, we consider a dynamic vehicle routing problem where one additional customer arrives at a beforehand unknown location when the vehicles are already under way. Our objective is to maximize the probability that the additional customer can be integrated into one of the otherwise fixed tours without violating time constraints. This is achieved by letting the vehicles wait at suitable locations during their tours, thus influencing the position of the vehicles at the time when the new customer arrives. For the cases of one and two vehicles, we derive theoretical results about the best waiting strategies. The general problem is shown to be NP-complete. Several deterministic waiting strategies and an evolutionary algorithm to optimize the waiting strategy are proposed and compared empirically. It is demonstrated that a proper waiting strategy can significantly increase the probability of being able to service the additional customer, at the same time reducing the average detour to serve that customer.

Fourteen dissertations were submitted to the 1990–1991 Transportation Science Section dissertation prize competition. The first prize was awarded by Dr. Linos F. Frantzeskakis for “Dynamic Networks with Random Arc Capacities, with Application to the Stochastic Dynamic Vehicle Allocation Problem.” Three honorable mentions were also awarded. The recipients are listed in alphabetical order: Dr. Saad A. H. Al-Gadhi for “Characterization of Crowd Behavior and Movement,” Dr. Rina Schneur for “Scaling Algorithms for Multicommodity Flow Problems and Network Flow Problems with Side Constraints,” and Dr. Garrett van Ryzin for “Stochastic and Dynamic Vehicle Routing in Euclidean Service Regions.” The abstracts of all 14 theses submitted are reproduced.

The Stochastic Dynamic Vehicle Allocation problem involves managing a fleet of vehicles over time in an uncertain demand environment to maximize expected total profits. The problem is formulated as a Stochastic Programming problem. A new heuristic algorithm is developed and is contrasted to various deterministic approximations. The paper presents computational results that were obtained by employing a Rolling Horizon Procedure to simulate the operation of the truckload carrier. Results indicate the superiority of the new algorithm over other approaches tested.