On the set of indices of convergence for reducible matrices (Q2277546)

The authors obtain explicit formulas for the index set \(R_n\) of the class of \(n\times n\) reducible Boolean matrices and the index set \(BI_n\) of the class of all \(n\times n\) Boolean matrices. Let \(E_n=\{k\mid k=\gamma (A)\) for some \(n\times n\) primitive matrix \(A\}\). The main theorems of the paper are as follows: Theorem 1. Let \(R_n\) be the index set of the class of \(n\times n\) reducible Boolean matrices (or the class of nonstrong digraphs of order \(n)\). Then \[ \cup^{n-1}_{i=1}\cup^{i}_{j=0}(E_{n-i}+j)\subseteq R_n. \] Theorem 2. (The expression for the index set \(R_n)\). \[ R_n = (\cup^{n-1}_{i=1}\cup^{i}_{j=0}(E_{n-i}+j))\cup \{0\}\quad (n\geq 2). \] Theorem 3. Let \(BI_n\) be the index set for the class of all \(n\times n\) Boolean matrices. Then \[ BI_n=R_ n\cup E_n=(\cup^{n-1}_{i=1}\cup^{i}_{j=0}(E_{n-i}+j))\cup \{0\}\quad (n\geq 2). \]