Robust Network Design and Last-mile Delivery in Cold Chain Logistics System

Abstract

As a new logistics system, cold chain logistics, varying greatly from conventional logistics mode due to the intense sensitivity to time of the fresh product, has attracted significant attention recently. To reduce the losses caused by long transportation time, people have put forward higher management requirements for the cold chain logistics circulation links, which, however, is challenging to meet considering the uncertainty of demand. Therefore, to ensure fresh product quality and improve customer satisfaction, a specific robust network designing problem (RNDP) and last-mile delivery (LMD) technical framework based on a case study is proposed in this paper. Firstly, we apply a two-stage robust model to reasonably locate the warehouse, determine the delivery quantity under the condition of uncertain demands. We adopt two algorithms, Benders decomposition (BD) and column & constraint generation (C&CG), to solve the problem. Then, a vehicle routing problem with soft time window (VRPSTW) is established and solved by variable neighborhood search (VNS) heuristic algorithm to make distribution decisions. What's more, a series of sensitivity analyses are also conducted to study the influence of crucial parameters on decision-making.