Location And Pricing Of Last Mile Delivery Based On Customer Choice Equilibrium

The rapid development of e-commerce has created new consumer demand,but at the same time it has increased the delivery pressure of express.In the era of Internet plus customers require the express delivery to achieve personalization and diversification.And the traditional single home delivery mode can not solve the problems such as low delivery efficiency,high delivery cost and low customer satisfaction.Thus the formation of a diversified last mile delivery service system is an important problem to be solved urgently in the express delivery industry.It includss how to designs reasonable pickup service and improve the quality of home delivery service.As the only phase of direct contact with the final customers,last mile express delivery has gradually become a key factor affecting the online shopping experience of consumers.Facing different delivery modes,customers often show bounded rational choice behavior,or rely on the traditional home delivery mode,or choose a single pickup mode,or select multiple delivery modes.In order to improve the customer's experience of the existing delivery services and solve the bottleneck of last mile delivery,the status of last mile express delivey modes and the decision making behavior of the customers were analyzed.Then,from the perspective of bounded rational customers,the quantal response equilibrium problem of different customer choice behavior when customers chose different last mile delivery modes was studied and two kinds of customer choice equilibria were discussed.Finally,according to the different decision-making enterprises,considering the different choices equilibrium behavior,the location of pickup points and the pricing of home delivery service are studied.Firstly,the paper analyzes the status of last mile express delivey modes and the decision making behavior of the customers.Through the literature review and the actual investigation,the classification standards of home delivery mode and pickup mode are put forward,and the problems in the operation process of existing delivery modes are revealed.The application scope of different delivery modes is compared and analyzed.At the same time,the customer's decision making behaviors including the customer demand characteristics,the decision-making factors and the decision-making process are analyzed.It reveals that the customer is not completely rational in the process of decision making,has the bias of decision making and shows the bounded rationality behavior of stochastic choice.Secondly,the quantal response equilibrium problem of different customer choice behavior is studied.Aiming at the correlational problem of different delivery modes,we take the customer selection between home delivery,attended Collection and Delivery Point(CDP)mode and unattended CDP mode as an example.The home delivery is simulated as an M/D/1 queue and the pickup point as differernt M/M/K/K queues.Nested Logit(NL)model is used to calculate the utility function of the pickup mode,and the customer choice models of three delivery modes are constructed.Then,the existence and uniquess of Nested Logit-Quantal Response Equilibrium(NL-QRE)in last mile delivery service system are proved.Aiming at the customer dependency problem of the home delivery mode,we take the customer selection between attended CDP mode and unattended CDP mode as an example.The pickup points are modeded as different queues.The utility function of pickup mode is modified by prospect theory(PT),and the customer choice models of two delivery modes are constructed.Then,the existence and uniquess of Prospect Theory-Quantal Response Equilibrium(PT-QRE)in pickup service system are proved.And the difference between the PT-QRE and completely rational choice equilibrium is analyzed theoretically.Numerical experiments verify the correctness of the choice equilibrium model,and reveal the degree of customer rationality and the degree of dependence on home delivery affect the design of the last mile delivery mode.Thirdly,the location problem of pickup points based on customer choice equilibrium is studied.We analyze whether the probabilistic-choice and optimalchoice choice behavior has an impact on the pickup point location.To improve the pickup point's operating efficiency and benefit,the mult-objective optimization model based on NL-QRE is formulated.Faced with the home delivery mode,the attended CDP mode and the unattended CDP mode,customers show bounded rational behavior which could not accurately assess the home delivery waiting utility or pickup loss utility.The non-dominated sorting genetic algorithm II(NSGA-II)is developed to solve the established optimization model,and compared with the weighted method and the ideal point.To meet the different interests of the two decision-makers in the distribution enterprise and the customer,the bi-level optimization model for pickup points location is proposed.Customers have reference dependence behavior in the home delivery mode and are lack of accurate calculaiton capability to asses loss value of the pickup mode.An iteration algorithm is designed to solve the upper and lower model based on NSGA-II and immune algorithm.The results verify the validity and feasibility of the models and algorithms,and show the customer bounded rational behavior and the home delivery price affect the pickup point network' operating efficiency and benefit.Finally,the pricing problem of home delivey based on customer choice equilibrium is studied.In the context of the fierce competition from the pickup service enterprises which service is free,the customer packages are easily rejected by the pickup points.In order to improve the revenue of the home delivery enterprises,the home delivery pricing model based on multiple delivery modes affecting each other is formulated.Aiming at the pricing model based on NL-QRE,the projection gradient method,the genetic algorithm and the sensitivity analysis based local search algorithm are designed,respectively.The results show that there are significant differences in the pricing strategies of different types of customers.Aiming at the pricing model based on PT-QRE,the improved projection gradient method is designed and compared with the genetic algorithm and the multistart local search algorithm.The results show that both the QRE and the PT-QRE affect the pricing model,and it is necessary for the enterprise to understand thoroughly the dependency of the home delivery mode and the degree of customer rationality.