Decomposition of a class of functionals

From MaRDI portal





Let \(({\mathcal E},{\mathcal F})\) be a Dirichlet space on \(L^ 2(X;m)\) and \((X(t),P_ x)\) be its associated Hunt process. It is well-known that for any \(u\in{\mathcal F}\), the additive functional (AF) \(u(X(t))-u(X(0))\) is decomposed to the sum of martingale AF and a continuous AF of zero energy [\textit{M. Fukushima}, Dirichlet forms and Markov processes (1980; Zbl 0422.31007)]. In this paper, the author gives a condition for the possibility of a similar decomposition of the functional \(g(X(t),t)- g(X(0),0)\). As in [\textit{T. J. Lyons} and \textit{W. A. Zheng}, Proc. R. Soc. Edinb., Sect. A 115, No. 3/4, 231-242 (1990; Zbl 0715.60096)], an adapted process \(A_ t\) is called a regular 0-quadratic variation process if \(A_ 0=0\) and, for any sequence of subdivisions \(\tau^ m=(t^ m_ k)_{k\geq 1}\) of \([0,T]\), \[ \sum_ n E_ m(A(t^ m_{n+1})-A(t^ m_ n))^ 2\to 0\text{ as }|\tau^ m|\to 0. \] A condition for \(f(X(t),t)\) to be a regular 0-quadratic variation process is given as \({\mathcal E}_ 1(f(\cdot,t))\in L^ 2((0,\infty))\) and uniformly continuous on each compact \(t\)-interval. Moreover, under the condition that \(\sum_ n{\mathcal E}_ 1(g(\cdot,t^ m_{n+1})-g(\cdot,t^ m_ n))\to 0\) and \(\sup_ m\sum_ n{\mathcal E}_ 1(g(\cdot,t^ m_{n+1})- g(\cdot,t^ m_ n))^{1/2}<\infty\), \(g(X(t),t)\) is decomposed to a martingale AF and regular 0-quadratic variation process.




Cited in
(23)








This page was built for publication: Decomposition of a class of functionals

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3983524)