Research On Electric Vehicles Transportation Scheduling And Charging Planning

As oil energy shortages and pollution problems become more and more serious,electric vehicles(EVs)receive social attention because of environmental protection and energy saving.Using EVs to provide ride-sharing services can not only reduce air pollution and energy consumption,but also improve vehicle utilization and alleviate traffic jams.In order to provide long-term services,it is necessary to study the transportation and charging scheduling of EVs.In this thesis,first,we consider the problem of transportation and charging scheduling of EVs under static information.Static information refers to the assumption that user travel demand data and real-time electricity prices are already known.We divide the solution into three steps.The first step is to allocate travel requests into ride-sharing set based on the insertion algorithm.Second,we use Integer Problem(IP)to model the charging problem,which aims to minimizing charging cost under the constraints of ensuring the power required and transport needs.In the last step,the principle of fair distribution is used to allocate transportation or charging tasks to each vehicle.It aims to ensure that the benefits obtained by each vehicle are fair.Afterwards,we study the problem that the travel demand data and the real-time electricity prices often change in real time.First of all,for the unknown problem of travel demand,we analyze the historical travel data and use LSTM model to predict travel demand.On the other hand,for the real-time price changing problem,the piecewise linear relationship between real-time electricity price and electricity consumption is defined,the competition relationship between EVs and other users is modeled using game theory.And we propose an algorithm based on double oracle algorithm to solve the best strategies.Finally,we uses the actual travel data of New York City for experimental simulation.The feasibility of the model and algorithm is proved,and through comparative experiments it is proved that our division algorithm is more fair.Moreover,LSTM predict model achieve good results.And experiment shows that the best charging strategies can reduce the electricity cost.