Crowd-shipping problem with time windows, transshipment nodes, and delivery options

Abstract

This research introduces a new variant of the vehicle routing problem in the last-mile delivery process - namely, the Crowd-Shipping Problem with Time Windows, Transshipment Nodes, and Delivery Options (CSPTW-TN-DO). Two types of fleets (i.e., dedicated vehicles and occasional drivers) are available to serve three types of customers. Type 1 customers require a home delivery. The parcel of type 2 customers must be sent to the selected alternative delivery point (ADP). Type 3 customers have the flexibility to either receive their parcel at home or at the selected ADP. Dedicated vehicles are able to serve all types of customers, whereas occasional drivers only make home deliveries. The objective of CSPTW-TN-DO is to minimize the total distribution cost of employing both fleets. We formulate a Mixed Integer Nonlinear Programming (MINLP) model for the problem and solve the model by the commercial solver CPLEX after applying a linearization process. We also propose an Adaptive Large Neighborhood Search (ALNS) to solve a set of newly generated CSPTW-TN-DO instances. The computational results indicate that the proposed ALNS provides high-quality solutions. In addition, we show that the VRPTW with a primary objective of minimizing the total distribution cost is a special case of CSPTW-TN-DO, and that the proposed ALNS achieves comparative performance to the state-of-the-art algorithms for VRPTW. After analyzing several scenarios, we conclude that simultaneously considering occasional drivers, transshipment nodes, and delivery options offers a great opportunity for a last-mile delivery system to reduce its total distribution cost.