1: \begin{abstract}

2: 	\setlength{\parindent}{0mm}

3: 	\textbf{Abstract.} We show fundamental properties of the Markov semigroup of recently proposed MCMC algorithms based on piecewise-deterministic Markov processes (PDMPs) such as the Bouncy Particle Sampler, the Zig-Zag Process or the Randomized Hamiltonian Monte Carlo method.  Under assumptions typically satisfied in MCMC settings, we prove that PDMPs are Feller processes and the space of infinitely differentiable functions with compact support forms a core of their generator. As we illustrate via martingale problems and a simplified proof of the invariance of target distributions, these results provide a fundamental tool for the rigorous analysis of these algorithms.

4: \end{abstract}

5: