Abstract: Based on the fuzzy timed Petri net，which has time characteristics，a new network attack model is set up.In the model，mainly the both-branch fuzzy sets and reliability theory are introduced.This model uses the superior domain（or inferior domain） of the both-branch fuzzy sets to represent promoting factors（or hindering factors） of an attack respectively，carries on the both-branch fuzzy sets-based decisions to decide influencing effects on the places in a transition’s post-set after the transition（an attack behavior） is fired.In the model，an attacking process is abstracted as a complicated system that can be repaired.After random failure，attack states represented by places will enter into a repairable phase which has a fuzzy time interval，and resume to attacking states after the repairable phase.In this way，the model can conform more truly to the fact that an attacker always keeps an attack state to be favorable to his attack.Finally the attack route generating algorithm of the model and the reliability analysis of system are made，and a simulation example is made to show that the model has good performances.

Key words: network attack model, fuzzy timed Petri net, both-branch fuzzy sets-based decision, reliability

摘要： 以具有时间特性的模糊时间Petri网为研究方法建立了一种新的网络攻击模型。在模型中主要引入双枝模糊集理论和可靠性理论。利用双枝模糊集中的上域（下域）表示攻击的促进因素（阻碍因素），并在此基础上进行双枝模糊决策以决定变迁（攻击行为）激发对其后集中库所（攻击状态）的影响效果。该模型认为攻击过程是一个可修复的复杂系统，库所代表的攻击状态在随机失效后，便进入一个具有模糊时间区间的修复期，并在修复期后恢复攻击状态，这样做更符合攻击者总是维持攻击状态，使其利于攻击的事实。最后，对这种模型的攻击路径和系统的可靠性进行了分析，并以例子对模型进行了仿真验证。

关键词: 网络攻击模型, 模糊时间Petri网, 双枝模糊决策, 可靠性