Research On Optimal Location Of Charging Facilities For Urban Electric Taxi Based On GPU Parallel Computing

As a green vehicle,electric vehicles can basically achieve pollution-free operation in the process of operation.It is an effective measure to alleviate the energy crisis and solve the current climate problems in China.Therefore,it has been vigorously promoted by our government.As an important part of urban public transport,taxi is one of the significant media to promote the rapid development of electric vehicles.However,the short range and difficult charging of electric taxis are the main reasons that restrict their large-scale popularization.Under the condition that the battery performance of electric taxi has not made a great breakthrough,it is an urgent problem to plan the charging station of electric taxi rationally and alleviate the "mileage anxiety" of taxi drivers to promote urban electric taxi.On the basis of the existing charging station layout,this paper studies the optimal location of the new electric taxi charging station in the city.The main work of this paper includes:(1)This paper summarizes the theoretical and technological status of urban public facilities location.Taking the unmet charging demand of electric taxis and the fixed cost of new charging stations as objective functions,a multi-objective location model of new single charging stations for electric taxis is proposed,and the calculation method of its objective function is described.Based on this model,an optimal location strategy for new charging stations is designed.(2)An improved multi-objective particle swarm optimization(BMOPSO)algorithm is proposed to solve the location model of a new single charging station.According to the optimized location strategy,an optimized location algorithm for new charging stations is designed and implemented.(3)In order to solve the performance bottleneck problem that does not meet the charging demand calculation,a GPU-based parallel computing algorithm for unsatisfied charging demand is designed,and the algorithm is implemented by calling the CUDA computing architecture with the PyCUDA interface.(4)An application example of the optimal location model for electric vehicle charging station is analyzed.Monte Carlo simulation method is used to calculate the daily average maximum demand of electric taxi charging in a city,collect and preprocess multi-source data related to the location of electric taxi charging station in the city,and describe in detail the solving process and results of the optimized location model of new electric taxi charging station.The results show that the optimized location scheme of charging station proposed in this paper is feasible and flexible,taking into account the unmet charging demand and fixed cost investment.The purpose of this study is to explore the feasible solutions for optimizing the location and layout of electric taxi charging stations,and to provide assistant decisionmaking methods and means for relevant management departments or enterprises.