改善人民群众看病就医感受是当前医疗管理政策的重要目标之一。但是，目前大部分与入院等待过程相关的研究均忽略了患者的等待体验，而传统的预约调度和等待提示等相关的理论又不适用于入院过程。因此，本项目以改善患者入院等待过程的体验出发，基于新型的时间窗告知模式，从系统分析、策略分析和策略设计三个角度，运用马尔科夫决策、机器学习、决策优化和算法设计等理论和方法对由时间窗告知策略和准入控制策略耦合而成的入院等待策略开展系列研究。首先对静态策略下的时间窗告知系统展开分析，然后通过固定一类策略对另一类策略进行优化求解，最后以策略库结合策略搜索算法的形式选择最优的策略组合来求解耦合的入院策略优化问题。本项目的研究结果不仅可用于患者入院过程，还可用于如银行、政府等办事机构的排队系统，具有重要的学术意义和较高的社会应用价值。

Improving the public’s experience during medical service is an important aim of the current medical management policy. However, most existing inpatient admission related researches ignore patients’ wait experience and traditional queue related theory, such as appointment scheduling or delay announcement, are not suitable to deal with the inpatient admission process. Therefore, this project would focus on the wait during the inpatient admission process to develop a series of researches on admission policies. It is based on the new time-window announcement mode; is studied from aspects of system analysis, policy analysis and policy design; uses theories such as the Markov decision process, machine learning, decision optimization, algorithm design and so on; is coupled by time-window announcement policy and admission control policy. This project will first analyze the time-window system under static policies; then optimize one of the policies by fixing another one; and finally, solve the coupled admission policy by constructing a policy library and developing an efficient policy-search algorithm. Achievements of this project would be applied not only in inpatient admission background but also in queues such as banks and government departments, which means this project would gain significant contributions to both academia and applications.