Revisiting matrix squaring (Q1176519)

Given an \(n\times n\) complex valued matrix \(A\), the author proposes an algorithm to find its spectral radius \(\rho(A)\) and critical ratio \(\kappa(A)\), i.e. the ratio of the second eigenvalue versus the first in magnitude, based on matrix squaring. The reason why such algorithm is raised in place of the conventional power method lies on the fact that it derives an upper bound of \(\rho(A)\) independent of \(\kappa(A)\). However the matrix squaring has some drawbacks, among which is the computational complexity. The fast multiplication of matrices is assumed to apply.