\(p\)-variation of strong Markov processes. (Q1879814)

Let \(\xi_t,\) \(0\leq t\leq T,\) be a strong Markov process defined on some complete probability space and with values in a complete separable metric space \((X,\rho).\) For any \(h\in [0,T]\) and \(a>0\) consider the function \[ \alpha(h,a)=\sup\{P_{s,t}(x,\{y: \rho(x,y)\geq a\}):x\in X,\;0\leq s\leq t\leq(s+h)\wedge T\}, \] where \(P_{s,t}\) is the transition probability function of the process \(\xi_t.\) One says that the Markov process \(\xi_t,\) \(t\in[0,T],\) belongs to the class \(\mathcal M(\beta,\gamma),\) \(\beta\geq 1,\) \(\gamma>0,\) if and only if there exist constants \(a_0>0\) and \(K>0\) such that, for all \(h\in[0,T]\) and \(a\in(0,a_0],\) \(\alpha(h,a)\leq Kh^\beta\slash a^\gamma.\) For \(p\in(0,\infty)\) and a function \(f:[0,T]\to X\) its \(p\)-variation is defined as \[ v_p(f)= \sup \left\{\sum_{k=0}^{m-1}\rho(f(t_{k+1}),f(t_k))^p: 0=t_0<t_1<\ldots<t_m=T,\;m=1,2,\ldots \right\}. \] The main result of the paper is as follows. Let \(\xi_t,\) \(t\in[0,T],\) be a strong Markov process with values in complete separable metric space \((X,\rho).\) Suppose that \(\xi_t\) belongs to the class \(\mathcal M(\beta,\gamma).\) Then for any \(p>\gamma\slash\beta,\) the \(p\)-variation \(v_p(\xi)\) of \(\xi_t\) is finite almost surely.