如何综合利用多个区分特征得到区分优势更大的区分器一直是密码分析者有待解决的问题。在假设多个区分特征相互独立的条件下,给出了综合利用多个区分特征的最优联合区分器,给出了该区分器的区分优势的计算方法,证明了联合区分器利用的区分特征越多,在相同的数据复杂度的条件下,区分优势就越大;在相同的区分优势的条件下,区分攻击的数据复杂度就越小。若利用单个区分特征的区分攻击的数据复杂度是N,则联合利用k个具有相同优势的区分特征进行区分攻击的数据复杂度约为N/k。
It is an open problem to combine many distinguishing characters to construct a distinguisher with more advan- tage. An optimal combined distinguisher based on multiple independent distinguishing characters was proposed. The computing method of distinguishing advantage of this combined distinguisher was presented and the conclusion was given that the data complexity decreased on the condition of same distinguishing advantage, and the distinguishing ad- vantage increased on the condition of the same data complexity when more distinguishing characters were used in distin- guishing attack. Let N denote the data complexity of distinguisher with only one character, then the data complexity of combined distinguisher with k-independent characters was about N/k for special case that all distinguishing advantages of k-independent characters were equal.