On supercontractivity for Markov semigroups (Q2641557): Difference between revisions

Let \((E,{\mathcal E},\mu)\) be a probability space and let \((D,{\mathcal D}(D))\) be a symmetric Dirichlet form on \(L^2(\mu)\). The author gives some necessary and sufficient conditions such that the so called super log-Sobolev inequality holds, i.e., \[ \mu(f^2 \log f^2) \leq r D(f,f) + \beta(r), \quad r>0,\;f\in {\mathcal D}(D),\;\mu(f^2)=1, \tag \(*\) \] for some decreasing function \(\beta\) on \(]0,\infty[\) with \(\lim_{r\to 0^+}\beta(r)=\infty\). In particular, he obtains explicit formulae for one-dimensional diffusions and for birth-death processes defined by \((D,{\mathcal D}(D))\). Moreover, the author (1) presents some others necessary and sufficient conditions for \((*)\) in terms of isoperimetric inequalities, (2) proves that \((*)\) is equivalent to the compact embedding of the Sobolev space in the Orlicz space generated by the Young function \(N(x)= x^2 \log (1+x^2)\).