Abstract: The small parcel delivery industry is experiencing rapid growth, primarily driven by the growth of the e-commerce sector. In this study, we introduce logistic models for the delivery of small parcels to a set of service points (SPs), and we present effective methods for solving them. In the traditional delivery model, each recipient specifies a single location at which they wish to receive the parcel; however, when SPs are used, many recipients may have no strong preference among several locations, e.g., near the recipient’s home address, near the recipient’s office, or in the recipient’s favorite shopping mall. If some recipients are flexible and willing to provide the sender with more than one delivery location, it is possible to perform the delivery task at lower cost and within a shorter amount of time. Our solution methods are based on the concepts of the savings heuristic, the petal method and tabu search with a large neighborhood. An extensive numerical study is conducted to evaluate our solution methods and demonstrate the benefits of our model compared to the traditional nonflexible one. We also present a simulation study to demonstrate that our model can be adapted to a stochastic and dynamic environment. A more revolutionary approach is to use the SPs as drop-off, pickup, and intermediate storage locations. A parcel may be carried from its origin to its destination in several legs via several possible intermediate service points. Such a system constitutes a physical internet (PI) service network. The PI is a generalization of the current practice of using a hierarchical network, where a parcel can switch vehicles only in a large sorting facility (hub) and a service point is served by a single route. The PI service network topology presents an opportunity to improve the delivery process by reducing the total distance that parcels are carried while still exploiting the possibility of shipment consolidation. In addition, such a system may save a significant amount of resources that are associated (and tied for an extended period) with the construction and operation of a large sorting facility. In this study, we develop tools to design and operate a successful PI for parcel delivery – a parcel routing mechanism and a math heuristic for routing and scheduling of the vehicles that transfer the parcels in the network. The effectiveness of the proposed tools and the advantages of the PI are demonstrated by a simulation study.
About the speaker: Ido Orenstein is a technical program manager, leading implementation and development of data and AI projects in the tech, commerce, finance and legal sectors. In 2021, Ido received his PhD degree from the Department of Industrial Engineering at Tel Aviv University, and in 2022 continued his research in a postdoc at the University of Southampton. His research focused on network and route optimization in the field of urban and regional transportation planning. During his PhD studies, Ido was advised by Prof. Tal Raviv. Ido received his B.Sc. (2008) and M.Sc. (2010) degrees in information systems engineering from Ben Gurion University of the Negev. Before his PhD studies, Ido was an information systems analyst and a development team leader in Soreq Nuclear Center and in The Technological Unit of the Information Corps of the IDF.