A Dynamic VRP with Varying Transportation Costs and Its Solution Strategy

Abstract

A vehicle routing problem (VRP) is a well-known routing optimization problem. This work presents a capacitated VRP where transportation costs among the customers change over time, and constructs its non-linear integer program. It applies to vehicle distribution optimization problems under a dynamic traffic network. Then, a dynamic optimization strategy with an enhanced Variable Neighborhood Search (VNS) is proposed for addressing this problem. It contains two steps, i.e., initial scheduling and rescheduling. The former provides an initial solution for vehicles and the latter generates a new solution if the traffic information changes. Furthermore, to simulate the transportation costs in the dynamic traffic network, a dynamic simulator with adjustable frequency and amplitude is designed. Finally, experiments are conducted and the results show that the solution quality is improved by 2% ~ 6% over that static scheduling strategy.