利用动态贝叶斯网络实现人群聚集风险分析

发布时间：2018-07-05 06:50

  本文选题：公共安全 + 风险分析　； 参考：《中国安全科学学报》2017年07期

【摘要】：为动态探究影响人群聚集风险的主要因素及定量评估人群聚集风险,依据贝叶斯估计理论改进静态贝叶斯网络模型,建立动态贝叶斯网络模型。用该模型可根据实时采集数据计算后验参数,获取动态定量风险评估结果。利用所建的动态贝叶斯网络模型动态定量评估北京市某大型商业街区人群聚集风险。结果表明:该街区初始人群聚集拥堵概率为0.8×10~(-3),踩踏概率为7.6×10~(-6)。随着实时观测数据的引入,最终拥堵概率为2.4×10~(-3),踩踏概率为1.63×10~(-5),其中疏散不及时、疏散通道不畅、疏散标志不清等3个因素的相对重要度影响因子较大,是主要影响因素。实例中各基本事件的发生概率和相对影响因子动态变化,证明该模型有效。
[Abstract]:In order to explore the main factors that affect the crowd aggregation risk and evaluate the crowd aggregation risk quantitatively, the static Bayesian network model is improved and the dynamic Bayesian network model is established according to the Bayesian estimation theory. The dynamic quantitative risk assessment results can be obtained by using the model to calculate the posterior parameters according to the real-time data acquisition. The dynamic Bayesian network model is used to evaluate the crowd aggregation risk in a large commercial district in Beijing. The results show that the initial crowd congestion probability is 0.8 脳 10 ~ (-3) and the stampede probability is 7.6 脳 10 ~ (-6). With the introduction of real-time observation data, the final congestion probability is 2.4 脳 10 ~ (-3) and the stampede probability is 1.63 脳 10 ~ (-5). Among them, the relative importance factors of three factors, such as untimely evacuation, poor evacuation passage and unclear evacuation sign, are the main influencing factors. It is proved that the model is effective by the dynamic change of the occurrence probability and relative influence factors of each basic event in an example.
【作者单位】： 首都经济贸易大学安全与环境工程学院;
【基金】：国家自然科学基金资助(71471121)
【分类号】：C912.4
，

本文编号：2099359