Parcel delivery network optimization problem considering multiple hubs and consolidation of small-sized parcels

Abstract

With the growth in delivery demand owing to the expansion of e-commerce and COVID-19, parcel delivery network optimization has become a significant issue for courier companies. This study introduces a real-world parcel delivery network problem in which parcels are transported from origin to destination spokes via one or two hubs. The parcels are sorted in hubs for distribution to spokes. Regular- and small-sized parcels are considered in the optimization problem. Moreover, small-sized parcels can be consolidated into gunnysacks based on their destination for efficient handling. A mixed-integer programming (MIP) model, large neighborhood search (LNS), and matheuristic algorithm combining the two aforementioned approaches were developed. Additionally, 90 benchmark instances reflecting real-world problems were generated. The MIP, matheuristic algorithm, and LNS performed well in small-, medium-, and large-sized instances, respectively. The benchmark instances and source codes are accessible online.