Ultrametric matrices and induced Markov chains (Q1922544)

Let \(I\) be a finite set. A nonnegative symmetric matrix \((A_{ij})_{i,j\in I}\) is ultrametric if \(A_{ij}\geq \min\{A_{ik},A_{kj}\}\) for all \(i,j,k\in I\). Such matrices are proportional to the potential function \(\sum^\infty_{k=0} P^k\) of some substochastic matrix \(P\) which, in turn, defines a graph \(G^A=\{(i,j): P_{ij}>0\}\). The paper mainly deals with the analysis of \(G^A\) and its restriction to the set of `roots' \(R=\{i\in I:\sum_{j\in I}P_{ij}<1\}\).