Coordinated Vehicle Routing with Uncertain Demand

Numerical optimization methods have been developed and applied successfully to many deterministic variants of the so-called vehicle routing problem (VRP). Unfortunately, existing numerical methodologies are not as effective for planning and design problems when uncertainty is a significant issue. In view of this, this presentation will show how approximation models for large-scale uncertain VRP's can complement conventional optimization methods and allow for the exploration of a broader set of design and operating strategies than is currently possible. The presentation will consider vehicle routing problems where vehicles have a finite capacity and demand is uncertain, focusing on strategies that coordinate the actions of all vehicles in the fleet in real time as information becomes available.