Subgeometric ergodicity of strong Markov processes (Q558691)

The authors are concerned with subgeometric \(f\)-ergodicity of a strong Markov semigroup \((P^{t})_{t\geq 0}\;\)on a locally compact and separable metric space \(X\). They present sufficient conditions implying that \(\lim_{t\rightarrow \infty }r\left( t\right) \left\| P^{t}\left( x.\cdot \right) -\pi \left( \cdot \right) \right\| _{f}=0\) for \(\pi \)-almost all \(x\in X\). Here \(r=(r\left( t\right) )_{t\geq 0\text{ }}\)is a subgeometrically increasing rate function, \(f\geq 1\) is a Borel function on \(X\), and \(\pi ^{\text{ }}\)is the unique invariant probability measure associated with \((P^{t})_{t\geq 0}\). The main condition is expressed in terms of modulated moments of delayed return-times to a small set. Analogous to the subgeometric \(f\)-ergodicity criterion for discrete-time Markov chains, it extends earlier criteria implying different notions of stability for continuous-time Markov processes such as Harris recurrence, positive Harris recurrence, ergodicity and \(f\)-ergodicity. It is also derived a condition for polynomial ergodicity that is expressed in terms of inequality on the semigroup generator and is analogous to the so-called drift inequality in the discrete-time case. Applications to Langevin tempered diffusions in \(\mathbb{R}^{n}\) and storage models are discussed in some details.