A Heuristic for the Time-Dependent Vehicle Routing Problem with Time Windows
We consider the Time-Dependent Vehicle Routing Problem (tdvrptw), a generalization of the Vehicle Routing Problem with Time Windows (vrptw) where the travel time between any pair of clients can vary over the time. Its purpose is to better handle the dynamic nature of the travel time, especially in urban areas where traffic congestion can have a significant impact on the transportation. We propose a heuristic based on column generation and on Variable Neighborhood Descent (vnd) for solving the tdvrptw. Several neighborhoods are used to identify improving columns at each iteration of the column generation process. Those columns are then stored in a shared pool. In the same time, the integer master problem is solved and its solution is then improved by the vnd. Both total distance and number of vehicle criteria are considered. Numerical results are then presented to show the interest of our approach.
KeywordsLocal Search Column Generation Master Problem Vehicle Route Problem Speed Profile
- 4.Golden, B., Raghavan, S., Wasil, E. (eds.): Vehicle Routing Problem: Latest Advances and New Challenges. Operations Research Computer Science Interfaces, vol. 43. Springer, New York (2008)Google Scholar
- 10.Toth, P., Vigo, D. (eds.): The Vehicle Routing Problem. SIAM Monographs on Discrete Mathematics and Applications, vol. 9. SIAM, Philadelphia, PA (2002)Google Scholar