Asymptotic behavior of eigenvalues and random updating schemes (Q688863): Difference between revisions

Let \(Q(T)\) be an \(M\times M\) stochastic matrix whose entries defined upon a small positive parameter \(T\). For \((Q(T))_{ij}\) asymptotic to a form \(\exp(-{1\over T}U_{ij})\) when \(i\neq j\), an algorithm is developed to find the convergence rate of the eigenvalues of \(Q(T)\) as \(T\downarrow 0\).