An effective way to solve the parking difficulty is to utilize idle private parking slots as temporary parking supplies under the shared parking mode. A platform-based management approach can make reasonable use of parking slots by gathering the parking time windows of owners and consumers in advance. However, the actual parking time is difficult to be consistent with that registered on the platform. Owners and consumers may be unpunctual when arriving and departing. The parking unpunctuality is a stochastic factor and affects the allocation of parking slots. This paper focuses on how to model and optimize the parking unpunctuality in the shared parking slot allocation problem. We formulate the parking probability function of consumers and owners in different time slots and transform the stochastic programming model into an expectation model. It is found that in a certain time slot, the optimal number of idle parking slot reserved is related to the probability of unpunctuality, as well as the number of cars registered to arrive and depart in the neighboring time slots. The numerical experiments indicate that when considering the parking unpunctuality, the allocation is less compact; the satisfaction of consumers is decreased, especially when the supply is limited or the consumer demands are excessive. In the middle and later periods throughout the daytime, parking slots are more likely to be overloaded.

解决停车困难的一种有效方法是在共享停车模式下利用闲置的私人停车位作为临时停车用品。基于平台的管理方法可以通过预先收集所有者和消费者的停车时间窗口来合理利用停车位。但是，实际停车时间很难与平台上记录的时间一致。业主和消费者到达和离开时可能不守时。停车违章是一个随机因素，会影响停车位的分配。本文着重探讨如何在共享停车位分配问题中建模和优化停车违规性。我们制定了消费者和所有者在不同时间段的停车概率函数，并将随机规划模型转化为期望模型。发现在某个时隙中，所保留的空闲停车时隙的最佳数量与不守时的概率以及注册的到达和离开相邻时隙的轿厢的数量有关。数值实验表明，在考虑停车不准时的情况下，分配较为紧凑。消费者的满意度下降，尤其是在供应有限或消费者需求过多时。在白天的中后期，停车位更容易超载。数值实验表明，在考虑停车不准时的情况下，分配较为紧凑。消费者的满意度下降，尤其是在供应有限或消费者需求过多时。在白天的中后期，停车位更容易超载。数值实验表明，考虑停车违规性时，分配较为紧凑。消费者的满意度下降，尤其是在供应有限或消费者需求过多时。在白天的中后期，停车位更容易超载。