在复杂网络中,威胁模型结构庞大、行为复杂,不利于建模后的威胁分析。该文从实现的角度出发,针对一类利用C程序实现的威胁对象及威胁,在已有的威胁建模理论的基础上,基于代数系统理论提出威胁对象及威胁的代数化刻画框架。基于该框架,采用代数簇理论建立威胁行为相似度度量函数,通过矩阵理论及非线性约束求解理论进行函数求解,从而实现相似行为的代数化判定。最后,针对判定后的相似行为,基于并发系统等价关系构建威胁行为聚合规则,实现威胁模型优化,减少威胁分析复杂度优化。 In complex networks, the threat model has a large structure and complex behavior, which is not conducive to modeling threat analysis. From the perspective of realization, this paper proposes a new algebraic framework of threat objects and threats based on the algebraic system theory, based on the existing threat modeling theory for a class of threat objects and threats using C programs. Based on this framework, algebraic cluster theory is used to establish the measure of threat behavior similarity, which is solved by matrix theory and nonlinear constraint theory, so as to realize the algebraic determination of similar behavior. Finally, according to the similar behavior after the judgment, the threat behavior aggregation rules are constructed based on the equivalence relation of concurrent systems to optimize the threat model and reduce the complexity of threat analysis.