This paper addresses a dynamic vehicle routing problem related to courier mail services with explicit consideration of multiple objectives. This problem is modeled as a dynamic traveling repairman problem with time windows, and with three objectives considered: maximization of the number of serviced customers, minimization of customer waiting, and minimization of total travel time. A solution (simulation) framework, extending techniques developed for single-objective dynamic routing problems in the literature, is proposed to tackle this multiobjective problem. An efficient chained local search heuristic is embedded in this solution framework to solve the underlying static problem, the traveling salesman problem with time windows. Comprehensive simulation experiments conducted on data sets derived from benchmark problems show that the proposed algorithmic approach is able to provide high-quality solutions with limited computing effort. These experiments also indicate that the multiobjective model proposed in this study can offer much more attractive solutions to decision makers than single-objective models.
Abstract