Abstract: In a multidestination multiple parking lot system, multiple parking lots are available near a destination, and drivers going to different destinations can be served by the same parking lot. The allocation of the parking demand of each destination to each parking lot is called the parking allocation problem. This paper defines the parking allocation equilibrium principle: the drivers going to each destination always prefer the parking lot with the minimum cost (including parking fees, searching time in the parking lot and walking time, where the parking fee and parking space search time are positively related to the current saturation of the parking lot). The parking lot with the second lowest cost is selected only when the parking lot with the lowest cost is full. All unused parking lots have a higher cost than those used or that have been full. Considering the real-time dynamic characteristics of parking demand and the number of vacant spaces in each parking lot, a mathematical programming model equivalent to the equilibrium principle is proposed, the uniqueness of its solution is proved, and a solution algorithm is designed. The model is verified by two examples, and the changing rules of parking space occupancy in each parking lot near each destination are reproduced to provide a basis for dynamic parking charging and parking location planning.

Key words: full information environments, dynamic, parking allocation, equilibrium, mathematical programming

摘要： 在一个多目的地多停车场系统中,每个目的地附近有多个停车场可供选择,每个停车场也可供多个目的地的停车需求停车。每个目的地的停车需求在各停车场的停车量即为停车分配问题. 本文定义了停车量分配均衡原则:各目的地的驾驶员总是首选最低费用(包括在停车场内的停车费用以及停车时间、步行时间转换后的货币成本,其中停车费用及在停车场内的车位找寻时间都正相关于该停车场当前时刻停车饱和度)的停车场停车;由于停车场具有容量限制,当费用最低的停车场已满则再选择费用次低的停车场,以此类推;所有没被使用的停车场或者比被使用的停车场具有更高的费用或者停车位已满。考虑停车需求和各停车场内空余泊位数的实时动态特性,提出了与该均衡原则等价的数学规划模型,证明了其解的唯一性并设计了求解算法。通过两个算例对模型进行了验证,并再现了目的地附近各停车场内车位占用变化规律,从而为动态停车收费、停车选址规划等提供依据。

关键词: 动态, 停车分配, 均衡, 数学规划