Research On Dynamic Automatic Driving Vehicle Team Scheduling With Time Window Based On Cloud Plan

During the “ Twelfth Five-Year Plan ” period,the digitalization of transportation resources has been steadily improved,the level of intelligence has continued to develop,the development environment has been continuously optimized,and the national key operation vehicle networking control system has been established,which has promoted the development and construction of national public transportation platform and service of road freight vehicles.During the “Thirteenth Five-Year Plan” period,the level of intelligence in transportation industry will be upgraded and at the same time,the planning of automatic driving will be strengthened to encourage innovation in key technologies.The digital transformation is accelerating the construction of the city.In order to promote the automatic driving vehicles on the road,many companies have developed automatic driving systems.In order to put the systems into use,they must meet the data sharing needs,such as running time,parking position and so on.The car cloud plan combines different groups such as travellers and service companies.Travellers do not need to purchase automatic vehicles instead they send travel requests to the cloud planning system through their mobile devices.The cloud planning system combines the travel demands of travellers with the status of automatic driving vehicles to supply the service for the travellers.The traditional vehicle scheduling problems mostly focus on the deterministic and static problem models.However,the problem of delivery in automatic driving vehicle routing planning usually involves the uncertainty of the size of the traveller's order,the traveller's time window,and the location of the vehicle.Therefore,this dissertation studies the dynamic automatic driving vehicle scheduling problem with time window,which has strong scientific significance and engineering application value.The main research contents of the dissertation are described as follows.(1)A vehicle scheduling problem with time window(VSPTW)order model is established.In this dissertation,the dynamic model is transformed into a static model through time segmentation when the new customer order is accepted.This new order model is a good solution to the difficulties of dealing with dynamic problems.Based on the new order VSPTW model,this dissertation studies the dynamic random heavy load in the vehicle delivery problem,the service strategy under non-gradual order,the dynamic and random vehicle problem with the traveller's eagerness,and provides a reference on designing an ant colony search scheduler.(2)An ant colony algorithm for dynamic vehicle scheduling with time windows is designed.A hybrid solver based on ant colony optimization system search architecture is designed.In the ant colony algorithm,a new algorithm is constructed by combining local search heuristic algorithm with automatic driving vehicle route problem,which has strong flexibility in dealing with dynamic problems.At the same time,the virtual pheromone path created in dealing with dynamic vehicle routing problems is a good solution for solving vehicle problems.Finally,based on the three common traffic scenarios of free flow,mixed flow and blocked flow,this dissertation compares the ant colony algorithm with the constructed heuristic algorithm in JADE,and verifies the efficiency of the ant colony algorithm.(3)The implementation process of automatic driving vehicle scheduling with time window under cloud plan is designed.Based on the complexity and dynamics of the vehicle scheduling problem,this dissertation analyzes the vehicle scheduling requirements and development principles under the cloud plan,and builds the vehicle scheduling system framework.Besides,the implementation process of the automatic driving vehicle scheduling is designed.