Opitimization Of Pickup Service Network Under Customer Demand Heterogeneity

The booming development of e-commerce has driven consumption upgrades.Now when the Internet is widely popularized,consumer appeals are gradually diversified,and the heterogeneity of customer demand is becoming more obvious.Express delivery volu me is growing rapidly,and the vigorous consumption demand has brought great pressure to the terminal distribution.With the low delivery failure rate and high flexibility,the pickup mode has been accepted by e-commerce enterprises,terminal distribution enterprises and end-customers,and has become an effective way to solve the "last-mile" problem of parcel delivery.However,there are still some problems,such as low density of pickup points,single service function and low service level.In addition,re gional development imbalance in terminal distribution promotes the optimization of pickup service network and the innovation in form of pickup point.Therefore,based on the heterogeneity of customer demand and the difference of regional development,it is an important challenge for enterprises to establish adaptive service network and layout convenient pickup points,developing diversified service products and improving pickup service level.Firstly,a demand function that reflected heterogeneity was estab lished,and the existing pickup mode was analyzed.The form of pickup point and network architecture based on customer heterogeneity was put forward.It revealed the objective influence factor for customers choosing pickup,such as convenient service points,diversified service products and differentiated service level.Moreover,it proposed that the subjective influence factors for customers were the difference of demand category and quantity,customer behavior.Then,a customer utility function that consi dered the decision-making phase was set up.This paper summarized the existing types of pickup points,the main service products and the network architecture in typical enterprises,combining with the development trends presented in actual operations,and put forward the multi-type standardized pickup point and the personalized pickup points for markets with different development level.Secondly,the location problem of multi-type standardized pickup point considering customer level was studied.Aiming at t he problem of insufficient infrastructure construction in the late start-up area,it was proposed to increase the coverage of network and made scientific decision-making for the layout of multi-type pickup points.Based on the demand characteristics of cus tomers with different levels and the supply characteristics of pickup points with different types,a utility function considering travel distance and attraction was constructed.A dual-target location model with the aims of largest customer utility and the lowest construction cost was established.The PSO was designed to solve the problem,which was also compared with GA,and the validity of the model was verified by the case of the construction of pickup points in Xinjin county.The results also revealed t hat the customer structure,the maximum acceptable distance,the capacity and cost of the pickup points would affect the decision.Considering the competitive pressure brought by the increasing number of enterprises,the paper described the customer selection behavior by probability function,introducing the cooperative coverage model to depict the common influence of multiple pickup points to customers,and constructed the competitive location model with the maximum coverage as the target.The GA was chose to solve the problem,and the cooperative coverage model was compared with non-cooperative coverage model,which proved that the former had advantages in demand coverage,coverage cost and customer utility.Thirdly,this paper studied the service composit ion problem in personalized pickup point considering multiple demands.In order to cope with the demand for refined and diversified service products in the region where terminal distribution had reached a certain extent,based on the early layout of the pi ckup points,service composition was designed.The service products could be divided into core service products and supporting service products according to the correlation degree of terminal distribution.Different cost functions and utility functions wer e set up based on the products' characteristics.The service composition model was constructed.The NSGA-II with improved crossover and mutation probability calculation method was designed,and compared with PSO;the influences of customer demand quantity,customer's perception of supporting service and unit variable cost difference on the total cost of enterprise were analyzed.The service composition model under competitive environment was also set up,introducing genetic operator to improve the ICA,and comparing it with GA,which revealed that the lower limit of coverage satisfaction level and enterprise budget affected pickup service composition.Finally,the optimization of service level for pickup points with the differentiation strategy was studied.When pickup became more popular and daily,customers had made more requirements on the convenience,service surroundings and pickup economy.The location and service composition of pickup points determined the distance for customer receiving service,and a ffected the customer satisfaction evaluation together with the service level and price.The utility function of customer was established,which considered distance,service level and price,and taking the influence of reference point on customer perception in multi-cycle decision making into consideration.Aiming at multi-type standardized and personalized pickup point,the decision maker provided differentiated service for multi-level customer groups or individual customers with multiple demands,and built a service level optimization model with the goal of maximizing total profit.Improving GA with GSA,the service level and corresponding price in multi-cycle were provided and the main factors influencing the supply elements were analyzed.Considering the scenario when decision maker determined the service level different of pickup point for the regions with diverse development levels of terminal distribution,this paper constructed a multi-objective model,which took the benefit of both enterprise and cust omer and the requirement of balanced development in different regions into account.The weighted fuzzy method was adopted to translate the model with single target and the validity of the model and algorithm was verified by an example.Based on the regional development of terminal distribution,three key problems such as location of pickup points,composition of services and optimization of service level under multi-type standardized pickup points and personalized pickup points were solved,which can be use d to meet the needs of various service provided in pickup points and heterogeneous characteristic exhibited by customers.It is helpful to instruct enterprises to understand customer behavior rule,innovating service concept,improving enterprises' adaptab ility to demand change.It also enables leaders to scientifically grasp the influence of internal and external factors on enterprise decision-making,rationally allocating resource elements,improving the quality and benefit of pickup service,and establishing more efficient and flexible service network to promote the continuous and healthy development of the Express industry.In the future,the research can be deepened by combining new problems,new trends and new technologies that are constantly emerging in the development of terminal distribution.