Approximation of Hunt processes by multivariate Poisson processes (Q5937969)

If \({\mathbf M}=(X_t,P_x)\) is a Hunt process associated with a Dirichlet form, then it was shown by \textit{Z. M. Ma, M. Röckner} and \textit{T. S. Zhang} [Ann. Inst. Henri Poincaré, Probab. Stat. 34, No. 1, 1-22 (1998; Zbl 0898.31009)] that \(X_t\) is approximated by a sequence of multivariate Poisson processes. In this paper, for a given Hunt process which is not necessarily associated with a Dirichlet form, a similar approximation without exceptional starting point is shown. Instead of the capacity defined by Dirichlet form in the previous paper, probabilistically defined capacity is used. Convergence of the canonical Markov chains on a compact metric space which is obtained by a completion using a countable family of quasi-continuous functions is first shown and then used to obtain the convergence in the original space.