Ergodic properties of Markov maps inR d

Summary For domainsIR ^d (bounded or not) the notion of a Markov map fromI into itself is developed. It is shown that under a condition of Rényi type and the assumption that the map is Markov, any probability density tends inL ¹-norm to a unique invariant measure under the action of the Perron-Frobenius operatorP . The smoothness and ergodic properties of that invariant measure are studied. The paper generalizes results of Lasota and Yorke from dimension one to higher dimension.