Invariant measures of C^ 2-diffeomorphisms of Markov type in R^ d

From MaRDI portal
Publication:1207669





Let \(I\) be a domain in \(\mathbb{R}^ d\) (bounded or not). A transformation \(\varphi: I\to I\) is called \(C^ 2\)-diffeomorphism of Markov type iff there is an at most countable family \(\pi\) of subdomains of \(I\) which form a partition of \(I\) and satisfy the following conditions: (a) for any \(U\in\pi\), \(\varphi|_ U\) is \(C^ 2\)-diffeomorphism, and (b) for any \(U,V\in\pi\), if \(V\cap\varphi(U)\neq\emptyset\), then \(V\subset\varphi(U)\). If in addition \(\varphi\) satisfies the irreducible condition: (c) for any \(U,V\in\pi\), \(V\in\varphi^ j(U)\) for some \(j\), then it is called a Markov map. Earlier, it was shown by the author [Probab. Theory Relat. Fields 88, No. 4, 483-496 (1991; Zbl 0723.60089)] that if a Markov map \(\varphi\) is uniformly expanding, satisfies Rényi's condition and certain mild additional one, then any probabilistic density tends, under the action of the Frobenius-Perron operator \(P_ \varphi\), to \(g_ 0\) --- the unique \(P_ \varphi\)- invariant density. In this paper we study the reducible case. It is shown that there is a linear projection \(R\) in \(L^ 1\) which maps \(G=\{g\in L^ 1:g\geq 0,\;\| g\|=1\}\) onto \(\text{Fix}(P_ \varphi)=\{g\in G:P_ \varphi g=g\}\) and that \(\lim_{j\to\infty}P^ j_ \varphi g=Rg\) in \(L^ 1\), for all \(g\in G\), and uniformly, on each \(U\in\pi\), for certain \(g\) of \(G\). The method of the proof enables one to determine the number of all ergodic densities of \(\text{Fix}(P_ \varphi)\) in a rather effective way.











This page was built for publication: Invariant measures of \(C^ 2\)-diffeomorphisms of Markov type in \(\mathbb{R}^ d\)

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1207669)