E-commerce city logistics synchronization

Abstract

E-commerce city logistics (ECL) refers to intra-city logistical activities distributing goods sold by means of e-commerce. As the section directly linking end users, ECL has attracted much attention due to its substantial influences on customers’ satisfaction. Numerous efforts have been devoted to ECL also because it is the costliest, least efficient and most polluting section within the whole e-commerce supply chain. After all, ECL faces dual challenges of both e-commerce logistics and city logistics. Synchronization (sync) is emerging as an important constraint commonly encountered in ECL, requiring a group of logistical operations to be completed (almost) simultaneously, to reduce waiting, increase resource utilization, improve customer satisfaction, lessen uncertainty, etc. Nevertheless, ECL with sync requirements is rarely studied in existing literatures. This thesis hence fills the gap by proposing four typical ECL scenarios with sync requirements. The first scenario addresses delivery sync in ECL, where multiple deliveries for the same customer need to arrive simultaneously. Sync constraints are innovatively modelled as sliding time windows (STW), which is special type of time window that can slide earlier or later with window size fixed. Multiple deliveries must be made within the same STW. Results have shown that sync service and transportation cost can be well balanced via adopting STW. The second scenario deals with a multi-echelon ECL scenario with pooling sync, where products ready at city depots are transported to customers through a set of intermedium facilities called satellites. Such networks are commonly seen in nowadays e-commerce express logistics. Pooling sync is considered at the satellites so that products from different depots can be consolidated and delivered to the same customer. Numerical studies have revealed that even very small storage capacity could achieve pooling effects and save the number of vehicles required. The third scenario tackles a B2B ECL with delivery sync, where a set of business customers (retailers, restaurants, manufacturers, etc.) demand for synchronous deliveries of multiple commodities. A decentralized decision-making mechanism is developed based on O-VCG auction. Bidders (i.e. customers) offer bids which contain sync requirements, demands, and prices, and the auctioneer (i.e. 3PL company) chooses the allocation that maximizes its own profit. The proposed O-VCG auction realizes incentive compatibility on the buy side, approximately allocative efficiency, budget balance, and individual rationality. It is hence dominant for each customer to bid truthfully. The fourth scenario discusses about a B2B ECL with pickup sync, where multiple types of wastes at the same customer need to be collected synchronously. O-VCG auction is adopted to realize decentralized decision-making. Since vehicle routing problems (VRPs) are embedded as sub-problems and need to be solved for an exponential number of times depending on the bidder size, variable neighborhood search (VNS) is adopted to strategically guide the local searches. Results have shown that dynamic customer demands contribute greatly to 3PL’s profitability. Also, when more types of wastes need to be collected, customers’ utilities are harmed while 3PL surprisingly gains more.