The privileged permit service can be provided as an alternative to the conventional meter and reserved services in the off-street parking lots. In view of the unbalanced demand and the simplistic off-street parking lot management, this paper proposes a novel parking management problem for setting up and withdrawing the temporary permit-only policy. To optimize the access rule regarding uncertainty demand on the time of day and the utilization of the parking lot, a deep Q-learning (DQL) method is proposed to address the uncertainty and dimensionality in the framework of deep reinforcement learning (DRL). To replicate real-world demand pattern for training deep Q network, a short-term parking demand model is presented by integrating the long-short term memory neural network and multivariant Gaussian process. A case study is performed on urban parking lots on university campus. The numerical experiments of a rule-based strategy, a tabular Q-learning (TQL) method, and the proposed DQL method are conducted to justify the effectiveness of the proposed method. The proposed method outperforms the static (s, S) inventory policy by 65% and TQL with linear Q-value estimation by 15% in the total revenue. The sensitivity analyses show the DQL method is capable to handle capacity-reduced, demand-increased, and special-event scenarios while the comparable strategy underperforms the proposed method

中文翻译：

---

基于学习的最优动态特权停车许可政策方法


可以提供特权许可证服务作为街边停车场传统计价器和预约服务的替代方案。针对路边停车场需求不平衡和管理简单化的问题，本文提出了一种新的停车管理问题，用于建立和撤销临时许可证政策。为了优化有关一天中时间的不确定性需求和停车场利用率的访问规则，提出了一种深度Q学习（DQL）方法来解决深度强化学习（DRL）框架中的不确定性和维度。为了复制真实世界的需求模式来训练深度 Q 网络，通过集成长短期记忆神经网络和多元高斯过程提出了短期停车需求模型。对大学校园的城市停车场进行了案例研究。基于规则的策略、表格 Q 学习 (TQL) 方法和所提出的 DQL 方法的数值实验证明了所提出方法的有效性。所提出的方法在总收入方面优于静态 ( s , S ) 库存策略 65%，优于线性 Q 值估计的 TQL 15%。敏感性分析表明，DQL 方法能够处理容量减少、需求增加和特殊事件场景，而可比策略的表现不如所提出的方法