A first and last mile crowdshipping problem by smart lockers in city logistics

  • 李 寶玉

Student thesis: Doctoral Thesis


As e-commerce grows and evolves a large volume of small freights and home deliveries need to be handled every day in city logistics A major challenging city logistics management problem for the couriers is to deal with failed parcel collection or deliveries This is the “first and last mile delivery problem” and may very likely lead to multiple ineffective receiving and delivery attempts with high logistics costs Here we propose a new shipping framework that utilizes pickup and delivery via smart lockers typically located in high traffic areas This network of smart lockers can serve as convenient access points to remedy the first and last mile delivery problem In addition to serving the purpose for directly delivering or receiving goods smart lockers can also be used for transhipment which in fact has almost been ignored in literature and in practice This function of transhipment will be very useful for crowdshipping companies since thousands of crowdsourced drivers are commuting between home and businesses with spare space in their cars in order to reduce shipping costs and effort shipping companies are also considering paying these independent drivers to deliver parcels for them on the way to their destinations In this research we investigate the multi-hop rider matching problem which takes the transhipment on nodes of smart lockers into consideration The objective of this problem is to minimize the total cost of delivering all the parcels under consideration on time We proposed two integer linear programming models: the Arc-based model and the Time-space network model Since these mathematical programming models are too time-consuming we propose a greedy algorithm focusing more on better directing shipping We also provide a rolling horizon algorithm to split the planning horizon into multiple smaller time periods and solve them sequentially From our computational experiments the results indicate that the initial solution obtained by our greedy algorithm does help to shorten the solution time of the integer programming models to some extent for solving cases of medium scale and the rolling horizon algorithm can help solve networks of larger scale in a shorter time
Date of Award2019
Original languageEnglish
SupervisorI-Lin Wang (Supervisor)

Cite this