Collaborative Distribution Route Optimization Of Truck And Subway Based On Data-driven

With the vigorous development of e-commerce and the acceleration process of urbanization,the problem of traffic congestion caused by the increasing demand of urban logistics and the shortage of public transportation resources has become increasingly prominent.At the same time,the requirements of urban residents for high-quality life urgently need a new urban logistics distribution model with efficient,fast and less impact.In order to save costs,reduce the occupation of public transportation resources by urban logistics,and protect the environment,a logistics model that uses trucks or other ground transportation vehicles conbine to subways to complete the distribution collaboratively is proposed to optimize the distribution process and provide a theoretical basis for distribution practice.In the process of coordinated delivery of trucks and subways,goods can be directly delivered by trucks after being collected by trucks,or they can be delivered by means of relay transportation between subways and trucks.Aiming at optimizing issues such as the truck collection and delivery routes,whether the subway participates in the distribution,and which customers are the transfer stations responsible for,a truck and subway coordinated delivery route optimization model is established that takes into account the uncertainty transportation time of truck and the schedule restrictions of subway.In order to take into account economic benefits and social benefits,the total delivery cost and delivery time are the optimization objectives.Aiming at the difficulty of solving problems such as the large scale of the problem in the optimization model,the mutual influence of uncertainty and schedule,the difficulty of solving the objective function,the expression of mutually exclusive optimization goals and the strong timeliness of the solution results,the ant colony algorithm is determined as the main body,the Monte Carlo method,Pareto theory and data-driven acceleration strategy and other auxiliary components collaborated to design an algorithm of multiobjective simu-ant colony optimization based on the data-driven strategy(DDS-MSAC)solve the problem.The validity verification and parameter optimization of multi-objective simulation ant colony optimization(MSAC)are completed by solving the calculation examples.The results show that the random forest is superior to Gaussian process regression,support vector regression and BP neural network in terms of the quality and time cost of the problem target value prediction.DDS-MSAC combined with data-driven acceleration strategies and MSAC optimization have no significant difference in results,but the search time is reduced by 50%.A collaborative delivery example was constructed for the Chongqing SF Express site and rail network to find the best delivery routes.Decision makers can choose the Pareto solution according to different decision-making demand.Compare and analyze the results of the truck-only delivery model,and find the truckmetro collaborative delivery model is more optimized than the truck-only delivery model in terms of total delivery cost,total delivery time,and truck driving distance,with optimization rates of 33%,22%,and 16%,respectively.Finally,the sensitivity of key parameters such as truck capacity and subway schedule interval are analyzed,and the relationship between parameters and optimization objectives is described to provide more information to help decision-making.