Escape of entropy for countable Markov shifts (Q2155192)

The authors study countable Markov shifts. For such systems they prove the upper semi-continuous nature of the entropy map and the fact that the ergodic measures constitute an entropy dense subset. Let \((\Sigma, \sigma)\) be a transitive countable Markov shift with finite topological entropy. It is proved that \[ \limsup_{n\to \infty}h_{\mu_{n}}(\sigma)\leq |\mu |h_{\mu /|\mu|}(\sigma) + (1 -|\mu |)\delta_{\infty}, \] where \((\mu_{n})_n\) is a sequence of \(\sigma\)-invariant probability measures converging on cylinders to \(\mu\). If the sequence converges on cylinders to the zero measure then the right-hand-side is understood as \(\delta_{\infty}\). The authors prove that the topological entropy of the system at infinity is equal to the measure theoretic entropy at infinity. In this non-compact scenario, a variant of Katok's entropy formula is obtained. For every transitive finite entropy countable Markov shift, they further show that the entropy map is upper semi-continuous.