基于贝叶斯网的舰艇防空威胁评估

Threat Assessment for Warship Air Defense Based on Bayesian Network

现有威胁评估模型的参数大多由专家直接给出,很难应对不确定的对抗场景,以舰艇防空为背景,探讨如何建立适应能力更强的评估模型。首先通过领域专家构建一个舰艇防空威胁评估的静态BN模型;接着加入时间变量,将威胁评估BN扩展为DBN模型;然后结合专家约束,用约束最大后验概率估计算法学习网络参数;最后,通过推理威胁评估的结果来验证模型的有效性以及对不确定对抗场景的适应能力。该建立模型的流程具有更好的普适性,能很好地处理不确定对抗场景下的威胁评估任务。

Most of the parameters of the existing threat assessment model are directly given by experts,and it is difficult to deal with uncertain confrontation scenarios.In this paper,the ship’s air defense is used as a background to discuss how to build a more adaptive assessment model.First,construct a static BN model of the ship’s air defense threat assessment by domain experts;then add the time variable to expand the threat assessment BN to the DBN model;then combine the expert constraints and learn the network parameters with the constrained maximum posterior probability estimation algorithm.The results of threat assessment verify the effectiveness of the model and its ability to adapt to uncertain confrontation scenarios.The model building process has better universality and can handle threat assessment tasks in uncertain confrontation scenarios well.