E-commerce Logistics Distribution Network Optimization And Simulation-based Algorithmic Development

With the rapid development of the digital economy and the gradual improvement of infrastructure,the demand for e-commerce logistics keeps growing,which leads to diversified logistics service products and makes the logistics supply challenging.How to provide rapid and efficient delivery services have attracted great attention from both research communities and many e-commerce logistics companies.Focusing on the challenge of city logistics distribution network,this thesis aims to provide a solid analytical foundation for decision-making by optimizing the underlying service network configuration and dispatching wave policies.Firstly,for city logistics distribution network that allows lateral-transshipments within distribution centers layer,the intra-layer flow is transformed into inter-layer flow through network reconstruction.Then a generalized quadratic assignment model with capacity constraints is developed.The model mainly makes decisions on the the allocation of warehouses to supply the corresponding distribution centers(DCs),and the demand coverage decision from distribution centers to delivery stations.These two parts of decision jointly determine the flow of lateral-transshipments,and the objective function is to minimize the overall transportation cost.Secondly,to solve the proposed model effectively,this thesis constructs a GLB-type lower bound by adopting thoughts of Gilmore and Lawler lower Bound(GLB)for QAP.Combined with this lower bound,both exact and heuristic solution algorithm are developed.To demonstrate the effectiveness of proposed algorithms,this thesis carried out computational experiments based on publicly accessible data sets and the real-world case,and compares the results with the Gurobi solver.The analysis results show the key role of GLB-type lower bound in improving the upper bound and the effectiveness of proposed algorithms.At the same time,it also shows the significance of lateral-transshipments in reducing the overall transportation cost.Thirdly,to further analyze the impact of dispatching waves on urban logistics distribution,the simulation network model is constructed.This thesis organizes main business processes of the entire logical path from the orders releasing to the fulfillment at delivery stations.Combined with the characteristics of dispatching waves and the business process of the whole logical path,this thesis provides a scalable simulation network modeling method,which integrates the business process node and the physical city distribution network.Finally,to optimize dispatching wave policies,this thesis designs and implements an agent-based city logistics simulation engine,and the efficient parallelism of the simulation process is realized through OpenMP framework and simulation event space division.Compared with the real-world data,the accuracy of the simulation is verified,and the relevant evaluation indexes are designed to evaluate the system performance.The simulation results under multiple scenarios show that the number of dispatching waves has a significant impact on the timeliness,and the selection of waves under different order quantities should be determined according to different cost and timeliness requirements.There are 32 figures,20 tables and 95 references in this thesis.