Motivated by the toll-by-weight schemes implemented in over twenty Chinese provinces, we propose a single vehicle routing problem in which the transportation cost per unit distance monotonically increases with the vehicle’s weight. This problem is a new variant of the traditional traveling salesman problem, and is a generalization of both the unweighted and weighted minimum latency problem. The branch-and-bound algorithm described in this paper is the first exact algorithm for the problem and can be implemented on instances with any monotonically increasing toll function. Experiments show that our algorithm outperforms the currently best known exact algorithms for the unweighted MLP and was able to find the optimal solution for an instance with 52 vertices; is also capable of solving weighted MLP instances of up to 52 vertices; and can find the optimal solutions for several SVRPTBW instances with up to 45 vertices.
The SVRPTBW is a good starting point for the investigation of the toll-by-weight scheme that is being used on Chinese expressways, and will continue to be used in the foreseeable future. The model does simplify certain aspects of the problem, and more complex models that consider additional factors can be fruitful avenues for future research. Possibilities include the multiple vehicle routing version of the problem, consideration of vehicle capacity, pickup as well as delivery, and time windows for delivery.