The maximal value for coefficients of ergodicity (Q1110917)

Let \(S_ n\) be the set of \(n\times n\) stochastic matrices, and \(\| \cdot \|\) be a vector norm on \({\mathbb{R}}^ n\). For \(P\in S_ n\), the ``coefficient of ergodicity'' \[ \tau (P)=\sup \{\| x'P\|:\quad x\in {\mathbb{R}}^ n,\quad \| x\| =1,\quad x'1=0\} \] is one way of measuring the worst possible rate of approach of \(x'P^ k\) to a limit as \(k\to \infty\), where \(x\in {\mathbb{R}}^ n\) is a probability vector and x' is the transpose of x. In this paper, the author finds explicit numerical expressions for \(\max \{\tau_ q(P):\) \(P\in S_ n\}\), where \(\tau_ q\) denotes the coefficient of ergodicity with respect to the \(\ell_ q\) norm, for each fixed \(q\in [1,\infty]\). These results unify previous work covering the cases \(q=1,\infty\).