| Since the beginning of the 21 st century,the number of cars in the city has risen sharply,resulting in the current parking facilities can not meet the parking demand,the problem of "parking difficulty" can be seen everywhere in the city,but with the rapid development of intelligent transportation system,parking reservation gradually attracted attention and application,become an effective means to alleviate the problem of "parking difficulty" in the city.Parking reservation before travel can help travelers quickly and accurately find the appropriate parking lot,avoid the situation of secondary parking,save travel costs,and relieve the pressure on the surrounding traffic system.Therefore,this paper studies the parking guidance decision based on reservation.Firstly,two different parking reservation modes are analyzed from the perspective of parking reservation.The real-time parking reservation mode is selected as the research object of this paper,and the specific parking process of the real-time parking reservation mode is constructed.On this basis,the grey model and Markov chain of short-term prediction methods are summarized,and the common methods and basic solution flow of multi-attribute decision making are introduced,which provides research ideas for the establishment of short-term prediction model of remaining parking space at reservation time and parking guidance decision model under reservation mode.Secondly,a questionnaire survey of parking reservation was carried out for travelers in the main urban area of Chongqing.Based on the questionnaire data,the demand of parking reservation was analyzed,and the feasibility of parking reservation was clarified.Based on the parking lot selection process of real-time parking reservation and related literature on parking induced decision making,the optimal parking lot decision attributes under parking reservation were analyzed,and five key parking lot decision attributes,namely,walking time,parking lot type,number of remaining parking Spaces,travel time and parking cost,were selected by fuzzy clustering method as the decision attributes of parking induced decision model.Then,the short-term prediction of the remaining parking space at the time of reservation was explored.GM(1,1)was used to construct the prediction model of the remaining parking space,and the Markov chain and metabolism methods were used to optimize the model,and the metabolism grey Markov prediction model was constructed.The model was used to predict and analyze the historical data of the parking lot,and the mean relative error was 1.71%.The fitting effect of the predicted and measured values is better than that of GM(1,1)model and grey Markov model,which indicates that the shortterm prediction of residual parking space is suitable and accurate.Finally,the multi-attribute decision-making problem of parking lot under parking reservation is studied,and the reliability analysis of parking reservation and the quantitative processing of parking lot decision attributes are completed.The initial screening of parking lot is carried out according to the remaining parking lot and the ideal radius of parking lot service at the time of parking reservation.The subjective and objective weights of decision attributes of primary parking lot are calculated by G1 method and entropy weight method respectively.Based on the ideal point method,the weight of the parking lot decision attribute combination was calculated,and the TOPSIS method was used to calculate the comprehensive attribute value of the primary parking lot set.The optimal parking lot recommendation under the parking reservation mode was completed.Finally,the parking lot decision model under the parking reservation mode was established,and the validity of the parking guidance decision model under the reservation mode was verified by taking a specific example of a user. |
