基于多智能体和元胞自动机人群疏散行为研究  

Study on Crowd Evacuation Behaviors Based on Multi-Agent and Cellular Automata Technology

作　　者：刘全平[1] 梁加红[1] 李猛[1] 付跃文[1] 

机构地区：[1]国防科学技术大学信息系统与管理学院,湖南长沙410073

出　　处：《计算机仿真》2014年第1期328-332,共5页Computer Simulation

基　　金：国家自然科学基金项目(61170160)

摘　　要：突发事件下人群疏散行为是安全领域的研究热点,分析复杂环境下行为决策和个人心理、生理对疏散行为和时间的影响,为公共场所安全设计和制订应急预案提供可靠依据。如何体现群体的整体性和个体的差异性是人群疏散仿真的难点。为解决上述问题,提出一种多智能体和元胞自动机的人群疏散模型。首先将元胞空间中被虚拟人个体占据的元胞视为一个独立的Agent,并将元胞及其状态进行封装,扩展为具有自主性的Agent。然后通过设计各种人群疏散行为策略做为演化规则,实现个体的差异性,较好地体现了个人个性、体力、心理(恐慌度)等对疏散行为的影响。仿真结果表明,考虑了个体内在因索的人群疏散模型更接近现实,进一步提高了模型对实践的指导意义。Urgent evacuation is the hotspot in safety field, and analyzing influence of evacuation behavior and time by behavior decision-making and individual psychology and psychology can provide credible basis in design of pubic building and making emergency plan. The difficult point of evacuation simulation is how model embodies crowd＇ s i- dentity and person＇ s individuality. For that purpose, we proposed a model of crowd evacuation based on multi-agent and cellular automata. In the proposed model, firstly each cellular occupied by one individual in the cellular space was considered as one independent agent. The cellular and its states were enclosed and extended as an autonomous a- gent. Then, a series of crowd evacuation behavior patterns were designed and individual differences were also taken into account. In this process, the influences of the person＇ s individuality, physical strength and psychology property have on the complex phenomena of crowd evacuation were reflected. Simulation results show that the proposed model considering the individual intrinsic factors reflects really crowd evacuation behavior in emergency situations and can give more guide to practice.

关 键 词：多智能体系统 元胞自动机 人群疏散 仿真 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]