Research On Parking Berth Sharing Scheme And Resource Dynamic Matching Model

With the acceleration of urbanization,urban land use is increasingly intensive,and the problem of parking is increasingly aggravated.Parking queuing and cruising aggravate the congestion of urban roads during peak hours.Vehicles in low-speed state produce a large number of harmful gases that threaten human health,pollute the urban environment,and bring serious consequences to individuals and society.Taking advantage of the spatial and temporal imbalance of parking demand,parking sharing among parking lots is considered as a key approach and effective means to improve the urban parking difficulty.However,in the process of practice,the effect of berth sharing is not ideal,and problems such as user demand conflict and low sharing willingness affect the sustainable development of berth sharing.Therefore,it is an important problem to be solved in the practice of berth sharing to determine the number and time of shared berths and to design a feasible dynamic matching scheme of berth resources.With berth sharing as the breakthrough point,the definition,basic conditions and participants of berth sharing are expounded.On this basis,the existing sharing mode and existing problems are analyzed,and the sharing platform reservation service framework is proposed.From the point of view of parking lot,quantitative analysis of parking lot indicators,qualitative analysis of parking lot sharing characteristics,to determine its sharing service scope and service objects.The time correlation of berth occupancy data is analyzed and the input structure of prediction model is determined.Based on LSTM neural network model,the parking space occupancy can be predicted in the future.According to the principle of sharing scheme,the scheme of the number of reserved berths and the method of the shortest sharing time are put forward.Combined with the error of prediction model,the number of shared berths and shared periods are determined.Starting from the problems existing in the matching process of shared berth resources,the price lever is adopted to adjust the attracting amount of shared demand and a dynamic matching model is integrated.Based on the binomial Logit model,the parking behavior model of drivers is established,which provides the basis for judging whether drivers choose shared parking lot in the dynamic matching model.Taking the maximum revenue of the parking lot as the objective function,and the shared time period and the data amount of the shared berth as the main constraints,the dynamic pricing berth resource matching model and the berth resource matching model considering the demand delay were established respectively.Shared berth utilization rate and shared demand rejection rate are selected as evaluation indexes of the model,and an algorithm is designed to solve the problem.Taking Changchun Eurasia New Life underground parking lot as the empirical analysis object,the validity of the model was verified.The results show that the prediction accuracy of LSTM model is high.The dynamic pricing berth resource matching model can significantly reduce the rejection rate of shared demand and increase the revenue of parking lots while maintaining a high utilization rate of shared berths.