A new algorithm for site-dependent vehicle routing problem

TLDR: A new algorithm for solving the site-dependent vehicle routing problem (SDVRP) is presented and applied to a set of 12 test problems taken from the literature and computational results show that the new algorithm easily outperforms the previous methods.

Abstract: In the site-dependent vehicle routing problem (SDVRP), a heterogeneous fleet of vehicles is used to service a set of customers, but there exist compatibility dependencies between customer sites and vehicle types. Some customers with extremely large demands may require large vehicles, whereas some customers located in congested areas may require small or medium-size vehicles. Other customers might be serviced by any type of vehicle. The goal is to carefully select an allowable vehicle type for each customer. Then, for each vehicle type, a classical vehicle routing problem (VRP) is solved over all customers selecting that vehicle type so that the total distance traveled by the entire fleet is minimized and all VRP constraints as well as the site-dependent constraints are satisfied. In this paper, a new algorithm for solving the SDVRP is presented and applied to a set of 12 test problems taken from the literature. The computational results show that the new algorithm easily outperforms the previous methods.