Exponential ergodicity for Markov processes with random switching (Q2345131)

The Markov processes considered in this paper are constructed from a Polish space \((E,d)\) and a finite set \(F\), a family \({({Z^{(n)}})_{n \in F}}\) of \(E\)-valued strong Markov processes, and a family \({(a( \cdot ,i,j))_{i,j \in F}}\) of non-negative functions on \(E\). The investigation is concerned with the process \({({X_t},{I_t})_{t \geqslant 0}}\), defined on \(E \times F\). The component \({X_t}\) behaves like \(Z_t^{{I_t}}\) as long as \(I\) does not jump. The process \(I\) is discrete and jumps at a rate given by \(a\). Imposing regularity assumptions on the jump rates and Wasserstein contraction conditions for the underlying dynamics, the authors ``provide a concrete criterion for the convergence to equilibrium in terms of [the] Wasserstein distance.'' In particular, they ``obtain exponential ergodicity in situations which do not verify any hypoellipticity assumption, but are not uniformly contracting either,'' as well as ``a bound in total variation distance under a suitable regularizing assumption.''