中文摘要：

对网络安全防御态势的有效预测，能够保障网络的安全稳定的运行状态。对安全防御态势预测，需要计算出当前的网络防御系数和态势变化值，来完成有效预测。传统方法融合D—S证据理论，得到安全可信度区间和相应防御系数，但忽略了对防御态势的变化趋势的分析，导致预测结果不准确。提出基于模糊动态贝叶斯网络安全防御态势预测模型。将模糊理论与叶贝斯理论相融合得到在当前防御态势下发生攻击的后验概率，统计网络历史防御的数据样本，组建历史网络安全防御态势下发生攻击时间序列，计算出当前的网络安全防御态势变化值、攻击发生概率、网络防御系数；组建网络安全防御态势的评估矩阵，采用遗传算法进行网络安全态势预测模型的优化，依据优化结果组建网络安全防御态势预测模型。仿真结果表明，所提模型预测精确度高，为保障网络安全稳定地运行提供了有力的依据。

英文摘要：

This paper proposes a prediction model for defense situation of network security based on fuzzy dynamic Bayesian theory. Firstly, the research integrated fuzzy theory with Bayesian theory to obtain posterior probability of at- tack under current defense situation and carried out statistics for data sample of historical network defense ,then built time series of attack taking place under historical defense situation of network security. Moreover, the research worked out variation value, attack happening probability and network defense coefficient of current defense situation. The e- valuation matrix of defense situation was built and genetic algorithm was used to optimize the prediction model. Final- ly, according to the optimization results, the prediction model was built. Simulation results show that the model has higher precision. It provides powerful gist for ensuring safe and stable operation network.