Study On Dynamic Pickup And Delivery Based On Rolling Time Domain In Instant Delivery Service

The dynamic pickup and delivery problem in the instant delivery service has the characteristics of uncertain order arrival,small product weight,and high delivery timeliness.It is quite different from the traditional logistics delivery with determined demand and is more complicated.In recent years,with the development of instant delivery services such as express delivery services,the instant delivery logistics industry has ushered in an explosion.It has a broad market scale and huge development potential.At the same time,with the continuous emergence of the Internet of Things,smart wearable technology,and the development of intelligent optimization algorithms with continuous development,it is feasible to study real-time order allocation and rider's route planning in real-time delivery.This paper takes the improvement of customer satisfaction with instant delivery services as the starting point and reduces customer waiting time in instant delivery as the goal.This paper proposes a solution to order allocation and route optimization in instant delivery,which has strong theoretical and practical significance.The dynamic pickup and delivery problem is different from the general static pickup and delivery problem.It is not possible to obtain all the order information before the solution process starts.The solution must be dynamically adjusted according to the real-time order request in the process of solving the problem.The solution is more difficult.Aiming at the characteristics of instant delivery,this paper proposes an order allocation and rider path planning model based on rolling time domain mechanism based on the current domestic and foreign research on the dynamic pickup and delivery of instant delivery,and designs a two-stage large neighborhood search algorithm solves the problem of dynamic delivery in instant delivery.The research content of this article includes the following aspects:(1)Analyze the characteristics of the dynamic pickup and delivery problem in real-time distribution logistics.Different from the traditional vehicle routing problem,each request for the dynamic pickup and delivery problem has two locations,the pickup point and the delivery point,and each has its own time window.The rider needs to visit the pick-up point and the delivery point successively within the specified time window,and the arrival time of the order is uncertain,and all the order information cannot be obtained before the problem is solved.On this basis,the dynamic pickup and delivery problem in instant distribution logistics is defined,and the optimization goal is to minimize the total waiting time of all customers without rejecting customer requests.Then,after analyzing the similarities and differences between static fetching and sending problems and dynamic fetching and sending,the real-time optimization methods for solving dynamic problems and the intelligent optimization algorithms for solving static problems are analyzed and compared.(2)Considering the characteristics and optimization goals of the dynamic pickup and delivery problem,model it.The hypothesis for solving the dynamic pickup and delivery problem is proposed,including the sequence of the rider visiting the pickup point and delivery point in the same order,and the rider cannot be replaced during the delivery of the same order,to ensure that the mathematical model established meets the constraints of the actual problem.(3)A two-stage heuristic algorithm based on the rolling time domain mechanism is proposed for the established dynamic fetching problem model.First,for dynamic problems,the rolling time domain mechanism is used to transform them into multiple static sub-problems.Then,on this basis,the greedy algorithm is used to construct the initial solution,and then the adaptive large-scale neighborhood search algorithm is used to iteratively optimize it,and Judge the acceptability of the solution through simulated annealing algorithm.(4)Design a test case using the real data of take-out orders from a certain platform in Dalian,and select two representative business districts as the source data of the case,and process them according to the model's assumptions,and finally generate 18 Group test case.(5)Design a comparative experiment to compare the algorithm proposed in this paper with other heuristic algorithms,and compare its performance in three dimensions of the number of overtime orders,the average order delivery time,and the cumulative overtime under different case sizes.The experimental results show that the adaptive large-scale neighborhood search algorithm combined with the rolling time domain mechanism proposed in this paper performs best in three dimensions,especially in large-scale examples.(6)According to the experimental results,it can be shown that the adaptive largescale neighborhood search algorithm proposed in this paper based on the rolling time domain mechanism is an effective algorithm to improve the quality of solving the dynamic pick-up and delivery problem in instant delivery,and can be compared Obtain high-quality solutions in a short time,which is suitable for real-time distribution fields that require high calculation time.