An efficient multistart heuristic for the driver and vehicle routing problem

This paper addresses the Driver and Vehicle Routing Problem, a complex routing problem with two depots in which the vehicles go from one depot to the other, while the drivers leaving from a depot must return to it within a given time limit. With these assumptions, it is mandatory for drivers to change vehicles in order to be able to go back to their base depots. The exchange of vehicles can only take place at some point. The objective is to design feasible routes for the vehicles and the drivers, so that the total cost is minimized. Only small instances of this problem can be solved to optimality. We present a multistart heuristic to tackle the problem in an efficient way. The computational experiments show that the proposed heuristic usually manages to find the optimal solution for the benchmark instances in the literature, being competitive when compared with previous heuristic results, and it is able to provide feasible solutions to much larger instances.