Research On Distribution Vehicle Scheduling Problem Based On Random Customer Requests

Vehicle scheduling problem is an important issue in logistics and distribution activities.Arranging a reasonable vehicle delivery route can effectively reduce logistics and distribution costs and improve service quality.In the actual logistics distribution process,due to the accelerated pace of people's life and the popularity of e-commerce,logistics distribution has been characterized by customer dispersion and random customer demand,which not only increases the difficulty of vehicle scheduling problems,but also reduces the economy of logistics distribution.In the process of vehicle distribution,when the customer demand appears randomly,in order to meet the customer's distribution needs,the distribution center will often change the vehicle's distribution route,or add distribution vehicles,which will reduce the economics of logistics and distribution.Therefore,how to effectively use the past customer information,timely respond to the random situation of customer demand in the distribution process,and improve the customer satisfaction while achieving the goal of optimizing the total distribution cost of logistics enterprises has become an important issue in logistics distribution vehicle scheduling problem.To this end,this paper will focus on the vehicle scheduling problem in which the customer demand appears randomly in the process of logistics and distribution.The main research contents are as follows:(1)Using the aggregate forecasting method to generate reasonable possible random location and demand of virtual customers by integrating stochastic demand information and empirical data,and the probability of occurrence;adding virtual customers to the actual vehicle dispatch with the actual customers who actually need the demand Formulation of the program.(2)Combining the virtual customer information and the probability of occurrence,the concept of weighted generalized total cost is proposed,and the customer satisfaction and the economics of the vehicle scheduling scheme are considered comprehensively,and the customer dissatisfaction is transformed into generalized cost,and the weighted generalized total cost is minimized.For the purpose,the vehicle scheduling model is established according to the distribution order of the real customer and the virtual customer.(3)According to the characteristics of the model,the improved genetic algorithm combining design and local search introduces adaptive operation to dynamically adjust the intersection and mutation probability,improve the efficiency of the algorithm,and quickly and efficiently generate the distribution vehicle scheduling scheme considering the virtual customer.(4)Through MATLAB to achieve improved genetic algorithm,select three sets of examples in the Solomon standard test study library,select three different sizes of customers,simulate the occurrence of two virtual customer requirements,and calculate the results of the study.The generalized total cost of the distribution of this paper is compared with the generalized total cost of the global rescheduling scheme,the additional vehicle scheme and the predictive scheduling scheme.The data results show that compared with the other three methods,the generalized total cost of the scheme is the lowest;and the scheme of this paper has universal applicability when solving the problem of different customer numbers and different customer location distribution forms.The vehicle scheduling method based on aggregate prediction proposed in this paper comprehensively considers the two factors of logistics distribution cost and customer satisfaction,can effectively predict before the random customer demand occurs and take it into account when formulating the vehicle scheduling plan,effectively reducing The scheduling cost when customer demand occurs randomly,improves customer satisfaction and service level,and provides decision support for logistics enterprises to respond to customer demand for random vehicle scheduling problems.