Research On Electric Vehicle Scheduling Strategy From The Perspective Of System Order

Because of its low carbon emission,electric vehicles have been more and more widely used in recent years.Due to a large number of electric vehicles are randomly connected to charging piles,causing the uneven distribution of charging demand.As a result,many electric vehicles spend a lot of time waiting in line,causing the lowly efficiency of charging system.In order to tackle this problem,this thesis models the two charging scenarios including single charging option and multi-charging options,and puts forward the corresponding scheduling algorithm to reduce the charging time of the system and improve the system order.In the single charging option scenerio,the goal is to reduce the system charging time and vehicle driving time.In order to solve this problem,this paper first models the charging scheduling decision-making problem as a Markov decision-making process,which schedules the electric vehicle by controlling the charging path and charging station selection of the electric vehicle.Because the decision space of the scheduling system is large,this thesis uses the neural network to approximate the value function,and solves the scheduling decision-making process by deep reinforcement learning.Simulation results show that compared with traditional scheduling algorithms,the proposed scheduling algorithm can significantly reduce the charging time of electric vehicles in the charging station by 33.9%.Futhermore,some of the charging stations can provide mulitiple charging options,which can influence the charging time for the charging system.In the multiple charging options scenerio,considering that the charging costs of different charging options varies,the aim should not only minimize the charging time and driving time,but also consider the charging cost.The model uses time of use charging price to regulate the user charging time to solve the charging conjestion problem.In order to solve the above problems,this thesis first consider the time of use pricing,and formulate the time of use price of charging according to the power load,so as to dispatch users to stagger peak charging.Secondly,in the multiple charging options scenerio,it is necessary to calculate the waiting time of electric vehicles in the charging station by queuing theory,and the waiting will be used in the decision-making of next scheduling.Finally,this thesis used the multi-objective optimization algorithm to solve the charging scheduling problem of multiple charging options,so as to reduce the charging time and charging cost.The simulation results first give a series of scheduling decisions,in which the charging time and charging cost are Pareto optimal;Secondly,the simulation experiment shows that compared with the fixed price,adopting time of use price can significantly reduce the charging time by 7.3%,and reuce the charging cost by 4%.