摘要： 针对基于Petri网攻击模型存在模型繁杂、规模大的缺点，提出从多条相关的命题规则生成多值逻辑Petri网(MVPN)的攻击模型。采用逆向推理简化模型，应用L-M算法对MVPN权值进行学习和训练，根据模型的性质与特点给出模糊逻辑推理算法。实验结果证明，该模型能对网络攻击行为进行描述，该算法能减小空间复杂度，提高计算效率。

关键词: 攻击模型, 多值逻辑Petri网, 权值学习

Abstract: Aiming at Petri net-based attack model existing the shortage of complicated model and large scale, this paper proposes a new complicated attack model based on Mutli-Valued Logic Petri-net(MVPN) by using a number of scattered but related rules. The backward reasoning method is used to simplify the complicated model, the back promulgation L-M algorithm is used to train the MVPN weights of the model, a fuzzy logic algorithm is put forward according to the nature and characteristics of the model. Experimental results prove that this model can clearly describe attack activities, its associated algorithm can reduce space complexity and improve the efficiency of calculation.

Key words: attack model, Multi-Value logic Petri Net(MVPN), weight learning

中图分类号: 

• N945.12