态势评估中的贝叶斯网络模型研究

针对战场态势评估的特点,采用模块化的贝叶斯网络模型对态势问题进行描述,动态构建贝叶斯网络结构.为了解决动态构建贝叶斯网络的推理问题,提出一种新的解决方法,将待连接贝叶斯网络模块的推理结果作为软证据,输入到已有网络对应的节点中,以代替所要连接的贝叶斯网络模块,在不改变贝叶斯网络结构的情况下完成网络的推理过程.仿真实验结果表明该方法是有效的.

Research on the Bayesian networks model in situation assessment

The modular Bayesian network is proposed, which is utilized to assess the situation in the battlefield. According to the event and target action, the modular Bayesian network is dynamically constructed. In the process of inference for the Bayesian model, the inference result of combined Bayesian networks is considered as the soft evidence of the corresponding node in the present networks. Finally, a simple application of the method is described. The results show that the method is available and computationally efficient.