 Geometric ergodicity

A Markov kernel is geometrically ergodic when its iterates converge to stationarity in total variation at a geometric rate, with a finite state-dependent prefactor.