Entropy formula and continuity of entropy for piecewise expanding maps (Q2664462)

Let \(\Omega\) be a compact subset of \(\mathbb{R}^{d}\), for some \(d\geq 1\). The authors obtain an entropy formula for an absolutely continuous invariant probability measure of a \(C^1\) piecewise expanding map \(\phi:\Omega\rightarrow \Omega\). Let \(\phi\) be a \(C^1\) piecewise expanding map with bounded distortion. It is proved that if \(\mu\) is a \(\phi\)-invariant probability measure such that \(H_{\mu}(\mathcal{R}_{\phi}) <\infty\) and the partition \(\mathcal{R}_{\phi}\) is quasi-Markovian with respect to \(\mu\), then \(h_{\mu}(\phi) \leq \int \log J_{\phi}d\mu\), where \(h_{\mu}(\phi)\) is the measure theoretic entropy of \(\phi\) with respect to the measure \(\mu\) and \(J_{\phi}\) is the Jacobian function. The authors also prove that if the partition \(\mathcal{R}_{\phi}\) is quasi-Markovian with respect to an absolutely continuous invariant probability measure \(\mu\) such that \(H_{\mu}(\mathcal{R}_{\phi}) <\infty\), then \(h_{\mu}(\phi) = \int \log J_{\phi}d\mu\). Then, they give an entropy formula for a \(C^1\) piecewise expanding map with bounded distortion with respect to an ergodic absolutely continuous invariant probability measure. Sufficient conditions for the continuity of the entropy of the absolutely continuous invariant probability measure are given. The authors prove that the entropy \(h_{\mu_t}(T_t)\) with respect to the absolutely continuous invariant probability measure \(\mu_t\) depends continuously on the parameter \(t\in [\tau,1]\). Lastly, they apply their results to a classical one-dimensional family of tent maps and a family of two-dimensional maps.