基于贝叶斯网络的作战目标评估

在作战指挥决策活动的筹划阶段,对敌体系要害和关键点的分析,当前尚未形成系统方法。针对此问题,利用贝叶斯网络在非精确知识表达与推理领域的优势,提出了综合考虑目标价值、打击难度、打击效果等因素的作战目标评估模型。根据判别贝叶斯网络分类器性能优于生成贝叶斯网络分类器的特点,在经相关领域专家论证的样本数据集的基础上,采取梯度下降法训练得出评估模型各结点条件概率分布。最后,利用Netica仿真软件,经样本数据测试,证明了作战目标评估模型的合理性。

Warfare Targets Assessment Based on Bayesian Network

During the operational command,there is no systematic method,in the analysis of the key points of the enemy system. Bayesian network has advantages in the field of imprecise knowledge representation and reasoning,which is just the value to establish an evaluation model. The model considers the target value,the difficulty and the effect. The performance of discriminative Bayesian network classifier is better than that of generated. So,the gradient descent method is used to carry out parameter learning on the basis of the sample data set which is demonstrated by experts. Finally,the sample data test proves the rationality of the model by the use of Netica.