基于马尔科夫链的停车寻位模型与仿真

A Markov-based Curbside Parking Model and Simulations

本文在经典的环形城市模型上,基于马尔科夫链建立解析的路边停车寻位模型.分别从系统和驾驶者两个角度,对寻位竞争队列进行描述,依据其马尔科夫特性,推导出停车寻位距离概率密度函数,发现传统的二项分布假设不能反映寻位车辆间的竞争车位行为,揭示了寻位车辆间的竞争是已有理论低估了停车难的原因之一.在数理分析基础上,提出环形城市路边停车仿真模型及算法,仿真结果验证了理论模型的结论,并发现车位被连续占用也是低估停车难问题的另一个主要原因.研究结果有助于加深对路边停车行为的认识,为制定相关车辆停车政策提供依据.

Referring to the classic circular city model, a curbside parking model based on the Markov chain theory is formulated. According to its Markov properties, the cruising competition queue is depicted through the viewpoints of the system and drivers, respectively. Consequently, the probability density function of the searching distance is derived, which proves that the traditional binomial distribution is limited in describing the competition behaviors between drivers. Then, a simulation algorithm is established to verify the theoretical results of the Markov model and find that besides the competition behavior, the continuouslyparking phenomenon is another major reason for difficult curbside parking. The finding extends our knowledge of curbside parking and can be helpful in designing the curbside parking policies.