中文摘要：

为更加科学准确地研究高铁车站在突发灾难下的应急救援策略，确保救援过程的安全性和实效性，以某高铁车站突发火灾事故为例，提出一种基于多Agent的救援仿真方法．首先，根据该车站交通疏散图建立灾难世界模型；然后，构建以环境Agent、待救援Agent、服务Agent、指挥中心Agent为主体的多Agent系统，并定义其属性和行为规则；第三，研究各Agent间的交互方式与通讯规则，建立彼此间交互组成复杂适应系统；最后，对救援过程进行仿真，并将多Agent的仿真结果与无救援、随机救援进行对比分析．研究结果表明：基于多Agent救援仿真系统制定的救援策略对高铁车站等结构复杂、空间密闭、人员聚集量大的场所突发灾难时，人员救助、扑灭火灾、清除路障等方面比现有的救援方式更为高效、安全；可实现理性智能的搜救和调度，为救援过程提供技术支持．

英文摘要：

In order to research the emergency rescue strategy of high-speed rail station under the catastrophic event scientifically and ensure the safety and effectiveness of rescue process, a kind of rescue simulation method based on multiple agent was put forward, with a high speed rail station emergency fire accident as an example. Firstly, according to the station traffic evacuation plan, a disaster world model was established. Then taking the environmental agent, waiting for rescue agent, services agent, command center agent as the main body, the multiple agent system are built, and its properties and behavior rules are defined. Thirdly, the interactive mode and communication rules among the agents are studied, and the complex adaptive system through the interaction between each other is established. Finally, the rescue process and compare the simulation results with the no relief and random aid analysis are simulated. The results show that the rescue strategies basing on multiple agent rescue simulation system for the high-speed rail station and other places of complicated structure, airtight space, a large quantity of persons gathered are more efficient and safe, when a disaster occurs suddenly, than the existing way of relief way in the assistance, extinguish fire, remove roadblocks and so on. And this study realizes the rational intelligent rescue and scheduling and provides technical support for the rescue process.