中文摘要：

为优化铁路客运专线车站的设计，研究了考虑不同发车频率的客流生成机制，应用标准的马尔可夫决策过程描述了乘客在车站内的集散过程，提出了基于逻辑网络和点阵的设备控制表示方法，建立了铁路客运专线车站乘客集散微观仿真模型。仿真发现：当行人移动空间在0．5m^2左右时，流量达到峰值，速度分布在20～80m·min^-1，并随密度增加呈指数递减趋势；在相同客流输入条件下，不同设施上的客流集散规律不同；通过对行人复杂行为的仿真，成功再现了客流集散中的自组织现象，有效避免了交通流基本图模拟现实交通中的不适应取值范围。结果表明模型具有较高的可靠性和实用性。

英文摘要：

In order to optimize the design of passenger dedicated railway station, the mechanism of passenger flow generation was studied, passenger mustering and evacuation process was described by using a standard Markov decision process, a method of facility control based on logical network and point matrix was presented, and a microscopic simulation model of passenger mustering and evacuation was proposed for passenger dedicated railway station. Simulation result shows that the maximum flow appears when passenger space is about 0. 5 m^2, the speed distributes between 20 m · min^-1 and 80 m · min^-1 and increases exponentially with density. Passenger patterns are different according to facility types despite of the same passenger flow input. The simulation of passenger behavior reproduces the self-organization phenomenon successfully, and avoids the inapplicable data sample range of traffic flow fundamental diagram. So the model is reliable and practical. 7 figs, 12 refs.