基于贝叶斯Noisy-or Gate网络对潜艇航行隐蔽性评估

Submarine Sailing Concealment Evaluation Based on Bayesian Noisy-or Gate Network

针对影响潜艇航行隐蔽性的部分典型因素特点,提出了基于贝叶斯Noisy-or Gate网络对潜艇航行隐蔽性评估的一种概率计算方法和网络模型,该贝叶斯网络利用专家知识分别从潜艇航行过程中的内因和外因分析,确定了主要风险因素,构造出有效的网络结构。对于没有考虑到的其他未知的因素,也会综合为一类因素加入网络中。采用满足Noisy-or Gate模型构造的网络节点参数,能使条件概率的个数减少,计算压力会大幅度地降低。在得不到充足样本数据时,该方法可以简化知识获取,还具有网络节点更新快速,实时性高的特点,可为潜艇隐蔽性航行提供定量评估依据,提高指挥员的决策效率。

Aiming at some specific factors affecting the concealment of submarine sailing,a probabilistic calculation method and network model based on Bayesian Noisy-or gate network for submarine sailing concealment evaluation is proposed.The Bayesian network using expert knowledge from submarine voyage are in the process of internal and external analysis,identified the main risk factors,construct an effective network structure.For other unknown factors that are not considered,they will be integrated into the network as a combination of factors.By using the network node parameters that satisfy the Noisy-or gate model structure,the number of conditional probabilities can be reduced,and the calculation pressure can be greatly reduced.When sufficient sample data is not available,the method can simplify knowledge acquisition,and has the characteristics of fast and real-time update of network nodes,which can provide quantitative evaluation basis for submarine concealed navigation and improve the decision-making efficiency of commanders.