Research On Time Window For Shared Parking Of Administrative Land In City Center

In recent years, the insufficiency of parking supply in the city center is more obvious as the speed of motorization. There are considerable parking spaces available in parking lots of administrative land when its own parking demand is at a low level. By sharing unused parking spaces of administrative land, limited space can serve more demand which is capable of easing parking problem. This paper sets parking lots of administrative land in city centers as the main study object, aimed at realizing the goal of making sufficient use of parking spaces through shared parking.Firstly, the composition of administrative land's parking demand is divided into three parts as private vehicle, official vehicle and visiting vehicle according to the investigation of typical administrative land. An analysis of parking distribution and features of cars in and out in weekdays and weekends is summarized. Parking demand due to meeting is also taken into consideration. Based on the above, shared parking adaption and models of administrative land are discussed.Secondly, a questionnaire survey of commuting users' and visitors' parking characteristic is conducted. This paper analyzes effects of different factors on commuting parking and parking behavior of visitors. Commuting users' and visitors' attitude towards shared parking is showed at the same time.Thirdly, this paper put forward a method, which combining wavelet analysis with BP-neural network, to predict the numbers of using parking spaces after the proof of predictability. Through a living example, the process of this method is shown with its effectiveness proved.Finally, on basis of several rules, different methods to split parking distribution of administrative land by time windows proposed during different periods such as working days, nights, weekends and meeting date. Under the constraint conditions of parking time and minimum numbers of shared parking spaces, procedure of these methods is performed though an instance. Based on the investigation of commuting parking, a binary-logit model is built to adjust the total numbers of commuting parking by changing parking price and administration's support policy.