Study On Demand Forecasting And Optimal Allocation Of Sharing Bike In Morning Peak

With the popularization of the green transportation,the impact of mobile payment and other emerging technologies,the sharing economy gradually challenges the traditional transportation industry.The shared bicycle market in China has experienced rounds of market changes since its first development in 2016.At present,the shared bicycle market in metropolis is experiencing a strong imbalance in the time and space of bicycle resources,which specifically shows the accumulation of vehicles at some points,and the phenomenon which does not provide enough vehicles available at other points,causing scheduling difficulties.This has seriously affected the sustainable development and refined management of the shared bicycle market.This paper accurately identifies the sever unbalanced area of shared bicycles and predicts the supply and demand conditions of vehicles during peak hours.Besides that,this paper proposes an optimal configuration model of shared bicycles in the region based on the combination of dynamic and static to maximize revenue based on the self-organization optimization concepts.The result shows that the model constructed in this research and the algorithm used to solve them have certain theoretical and practical value.First of all,the data preprocessing methods such as data cleaning,spatial matching,and data aggregation is used by the author to do the data cleaning.After that,the time and space characteristics of shared bicycle travel in the morning peak are statistically analyzed for the following research steps.The Louvain-based community network discovery algorithm was used to accurately identify the key areas of "Unbalanced areas" of shared bicycles in Xiamen City in the morning peak.Finally,122 clustering areas were obtained,and the top 40 areas were selected for key analysis according to the severity of "siltation".The results show that this area recognition algorithm that considers both geospatial constraints and the capacity of electronic fence sites for vehicles has good traffic significance and can obtain more accurate clustering results.Secondly,the CNN+LSTM method is used in this paper to predict the shared bicycles’ OD matrix during the peak period in Xiamen.Through projecting the preprocessed data into a geographic file for rasterization,adding time dimension on this basis,counting the OD numbers under different grids at 10 minute intervals as the input matrix,and use the previous 30 minutes of historical information to predict The traffic volume in the last 30 minutes is used to train the model with 80%as the training set and 20%as the test set.The results show that although the prediction result is interfered by a small number of "0" data in the matrix,the sigmoid function can quickly converge and have a stable performance under different time slices.The predicted OD data is processed by quantile processing.The results show that CNN+LSTM has accurate and efficient prediction accuracy for the OD change trend of the electronic fence in each area under the time slice in the future,which provides a certain guiding significance for the subsequent research on the optimal scheduling of shared bicycles.Finally,this paper designs a profit maximization optimization model based on the selforganization optimization concept.We comprehensively consider the benefits of operation scheduling,user convenience,and the benefits of user location distribution in the future,and establish a dynamic path planning model based on "initial static optimization+real-time dynamic optimization".The number of stops is optimized.The calculation results can be used to remind users through the mobile phone APP,assist user incentive measures,and actively guide parking users to park at nearby parking spots,perform peak shaving and valley filling,and alleviate the congestion problem of tidal spot parking spaces.Based on the case analysis of the Ruanjianyuan distract and Wushipu Metro Station distract,the results show that this optimization method can effectively solve the problem of imbalance between supply and demand of shared bicycles and has strong practical value.