Joint Optimization On Online Supermarket Order Processing And Vehicle Dispatching Based On Time-space Network

E-commerce has been leading the fast-developing internet economy in China,which has produced a number of large online supermarkets with Chinese characteristics.Large online supermarkets as a new e-commerce model,its convenient shopping modes and millions kinds of goods have brought great convenience to consumer life,but its own operation development also faces many challenges at the same time.Due to variable order time and large number of multi-item orders,online supermarkets need to allocate thousands of orders to warehouse every day.The unreasonable allocation of orders will not ouly cause the high logistics cost and delay receipt time,but also lead to decreased satisfaction.Thus,it is the key problem for the development of large online supermarket to process consumer orders scientifically and efficiently and shorten the package delivery cycle.This thesis takes online supermarket order fulfillment process optimization as the research background,and concentrates on order picking and package distribution.Meantime,we start with analyzing the inherent characteristics of warehouse and order and fulfilling "time dynamics","space complexity"and "coupling of decision variables" around online supermarket orders.Based on the timespace network methodology,particle swarm optimization and genetic algorithm,we joint optimization of order fulfillment and vehicle dispatch on online supermarkets.The thesis contains two aspects from the perspective of theoretical and practical.It has become an urgent problem to coordinate order picking and delivery in different zones with lower cost in the shortest time.Motivated by these issues,we study the problem of integrating order picking and vehicle dispatching.Firstly,we introduce a time-space network framework to model the order picking and delivery process,and formulate it as a mixed nonlinear integer-programming model.Then we adopt the hybrid particle swarm optimization algorithm to convert model,which introduced the crossover and mutation from the genetic algorithm to expand searching space and obtain a better fitness value.Next,we carry out numerical experiments to confirm the precision and accuracy of the algorithm by comparing with standard particle swarm optimization.Finally,we conduct sensitive analysis to generate managerial insights in practice.From the perspective of the last-mile package distribution of online supermarkets,we focus on the problem of low efficiency,high cost and frequency distribution,make comprehensive decisions about package distribution time,dispatching of vehicle.Firstly,we propose package consolidation(PC)approach consolidates package that arrive at different times for the same customer in the delivery station,and we formulate it as a mixed integer nonlinear programming model for package consolidation and vehicle dispatching problem to reduce logistics costs and achieve reasonable package allocation.Then,we propose adaptive particle swarm optimization algorithm with tabu search according to package consolidation rules(APSOTSPC),and demonstrate the precision and accuracy of the algorithm by comparing it with CPLEX.Finally,we conduct numerical experiments to demonstrate the superiority of PC approach and generate managerial insights for applying it in practice.This thesis provides a new solution for order processing in large online supermarkets and vehicle dispatching work.The order processing scheme proposed in this thesis provides a reasonable,efficient and scientific order fulfillment strategy for large online supermarket operators.In addition,it has enriched the theoretical research system for order fulfillment of large scale online supermarkets,and improved order fulfillment operations more scientific and practical in online supermarkets.