Abstract: Based on the advantages of probabilistic neural network in classification，a new decision-making method for evaluating damage of shocked naval vessel systems in the analysis of the survivability is presented，which regards evaluating damage of vessel systems in shock environment as a classification problem to solve.The model of classifying survivability of shocked naval vessel is established with limited training samples.As an illustrative instance，the vulnerability of a naval vessel’s subsystem is evaluated，which proves the validity for the application of PNN in the evaluation of vessel vulnerability.In numerical experiments the effect of relevant parameters on classification results are also discussed.

Key words: naval vessel, Probabilistic Neural Network（PNN）, classification, damage, vulnerability

摘要： 利用概率神经网络的分类功能，提出了一种新的易损性分析方法，该方法将受冲击舰艇易损性的评估问题作为一个分类问题进行处理，根据有限的学习样本建立了预测舰艇在冲击荷载作用下生命力的计算模型。对某舰艇主动力系统易损性的计算实例表明，该方法准确率较高。在数值实验中还分析了有关参数对系统损伤分类结果的影响。

关键词: 舰艇, 概率神经网络, 分类, 损伤, 易损性