Research On Vehicle Scheduling Method Of One-way Shared Electric Vehicle Based On Demand Forecasting

With the development of the automobile industry towards the trend of intelligence,sharing and electricization,one-way Shared electric vehicles have great development prospects in China.One-way Shared electric vehicles are not only economical and environmentally friendly,but also greatly improve the utilization rate of vehicles and parking Spaces due to the Shared mode of everyone,which is one of the effective methods to solve the problems of traffic congestion,parking difficulties and air pollution.However,due to the tidal and unbalanced travel demand of customers,the unbalance of vehicle inventory between sharing stations is hindered by the development of one-way Shared electric vehicle industry.However,in the existing studies,the distribution of target vehicles is mostly considered as given,which is less likely to be realized in reality.Therefore,this paper proposes a vehicle scheduling method based on demand prediction,which provides theoretical basis and scientific practical guidance for the development of one-way Shared automobile industry.Firstly,the vehicle demand forecasting problem of Shared site is studied.The vehicle demand of one-way Shared electric vehicle is introduced and analyzed.And the vehicle demand problems of Shared sites are classified,and the characteristics and influencing factors of each prediction problem are explored.Then,combined with the characteristics of vehicle demand forecasting,a vehicle demand forecasting model based on the combination of long-term memory neural network and radial basis neural network is designed.Secondly,based on demand prediction,vehicle scheduling is studied.The vehicle scheduling problem between stations is regarded as a kind of supply and demand distribution problem,the stations with vehicle backlog are regarded as supply stations,and the stations with vehicle shortage are regarded as demand stations.A multi-objective scheduling model with minimum scheduling cost,inventory cost and potential profit loss is established on the premise of satisfying constraint conditions.An improved simulated annealing algorithm vehicle scheduling model based on genetic algorithm is designed to solve the problem.Finally,in order to verify the effectiveness of the proposed method,Shanghai one-way Shared electric vehicle is taken as the research object to conduct an example analysis in this paper.The case study is introduced and the data used in the case is prepared.Then,the prediction model designed in this paper is used to predict the vehicle demand of the Shared site,and the prediction model proposed in this paper is compared with the long-and short-term memory neural network and radial basis neural network models respectively,which shows the characteristics of fast computing speed and high precision of this model.At last,on the basis of vehicle demand prediction,the vehicles of the Shared site are scheduled.The GA-SA algorithm designed in this paper is used to solve the scheduling problem.In order to verify the validity and rationality of the model and algorithm designed in this paper,this algorithm is compared with simulated annealing algorithm and genetic algorithm respectively.In addition,the scheduling method based on demand prediction proposed in this paper is compared with the ordinary scheduling method,and the results show that the method proposed in this paper is better,achieving a win-win situation between enterprises and customers.