The MDP Model And Application Of Order Allocation Decision Considering Stochastic Demand For The Same-day Delivery

With the massive application of emerging IT technologies such as big data,cloud computing and Internet of things in China’s express delivery industry,consumers have put forward higher requirements for offline logistics experience,thus giving birth to a new logistics delivery service model,namely "same-day delivery service".Same-day delivery service is a logistics service that directly realizes instant delivery from point to point,which directly meets the demand that goods can be delivered quickly and on time,and provides humanized upgrade service for modern people.However,due to the lack of an efficient scheduling method for sameday delivery service,the problems of low user satisfaction,high logistics cost and low distribution efficiency are caused.Faced with these problems,the platform should adopt an effective order allocation decision method for operation scheduling in order to reduce the delivery cost and improve the delivery efficiency.Therefore,it is of great significance to study the order allocation decision considering stochastic demand in same-day delivery service.On the basis of summarizing the research on order allocation decision at home and abroad,this paper makes an in-depth study on order allocation decision considering stochastic demand in same-day delivery service.From the previous research results,the researches on the order allocation decision problem mainly focus on vehic!e path planning and static vehicle scheduling,while the dynamic research on order allocation is rare.However,the researches on the selfowned platform distribution and the collaborative distribution between the self-owned platform and the third-party platform in this decision problem are more scarce.Therefore,this paper studied the order allocation decision problem with time window of self-owned platform distribution and the order allocation decision problem with co-distribution between self-owned platform and third-party platform,constructed the corresponding Markov decision process(MDP)model,and obtained the order allocation strategy under specific circumstances.The main research work of this paper is as follows:(1)The order allocation decision problem of the time window considering the delivery of the own platform in same-day delivery service is refined,and the order allocation decision is characterized by the markov decision process.In this paper,the order allocation decision problem considering time window of self-owned platform distribution is described,and according to the characteristics of the problem,a finite-dimensional continuous time markov decision process model is constructed to minimize the expected distribution cost of the platform.and the algorithm of solving the model is designed.(2)The order allocation decision problem of collaborative distribution between the own platform and the third-party platform in same-day delivery service is refined,and the order allocation decision is characterized by the markov decision process.In this paper,the order allocation decision problem of collaborative distribution between the self-owned platform and the third-party platform is described,and according to the characteristics of the problem,a finite-dimensional continuous time markov decision process model is constructed to minimize the expected distribution cost of the platform,and.the algorithm of solving the model is designed.(3)In this paper,the time window of self-owned platform distribution and the order allocation decision problem of co-distribution between self-owned platform and third-party platform are analyzed,and the optimal allocation strategy is obtained to verify the effectiveness of the proposed method.Based on the same day of service in the order of the stochastic demand allocation decision problem are considered in the different kinds of that day of service,have put forward their own platform distribution considering the time window and its platform with the third party platform distribution method,can effectively reduce the distribution costs,improve the efficiency of distribution,improve customer satisfaction and minimize the delivery cost of the platform.The research results of this paper provide reference and guidance for the further study on the dynamic decision problem of order allocation.