A class of strong limit theorems for inhomogeneous Markov chains indexed by a generalized Bethe tree on a generalized random selection system (Q1760002): Difference between revisions

A Markov chain is defined on the vertices of a generalized Bethe tree. Each vertex at level \(n\) (at distance \(n\) from the root) has the same number \(N(n+1)\) of edges to the next level, thus there are \(N(1)N(2)\dotsm N(n)\) vertices at level \(n\) for \(n=1,2,\dots \). Using terminology from gambling systems, a gain function is defined on pairs of states in a transition of the Markov chain, and a betting strategy is defined as a function of the states in the first \(n\) steps of the Markov chain. The total gain and the net gain in the first \(n\) steps are combined to a martingale process, and strong limit theorems are obtained by applying martingale convergence.