模糊离散动态贝叶斯网络的目标威胁等级评估

摘要

动态贝叶斯网络作为一种智能推理工具在处理不确定推理问题中显示出强大的生命力,但是存在难于处理连续变量的推理问题.将模糊理论与动态贝叶斯网络相结合,提出一种模糊分类的方法,将连续变量模糊分类为动态贝叶斯网络能够应用的证据信息用于推理,并建立目标威胁等级评估模型,应用直接推理算法对该网络进行推理.仿真结果表明,该分类方法与动态贝叶斯网络结合能够很好地处理连续变量推理的问题.%Dynamic Bayesian networks as a tool for intelligent reasoning problems show great vitality in dealing with uncertain reasoning, but there are some difficult to deal with the inference problem of continuous variables. This paper combines fuzzy theory and Dynamic Bayesian networks, puts forward a method of fuzzy classification to make the continuous variable into the evidence of Dynamic Bayesian networks, establishes the threaten grade assessment model of target and uses the directly reasoning algorithm for reasoning. Simulation results show that the classification method combining with Dynamic Bayesian network can well deal with the reasoning of the continuous variables.