Itô-Wiener chaos expansion with exact residual and correlation, variance inequalities

Consider a one-dimensional SDE: \[ dX_t= \sigma(X_t) dB_t+ b(X_t)dt,\;t\geq 0, \quad X_0= \xi\in\mathbb{R}. \tag{1} \] Assume that the coefficients \(\sigma,b: \mathbb{R}\to \mathbb{R}\) are in \(C_0^\infty\) and that the transition probability \(P_t(x,dy)\) associated with (1) exists and \(P_t f(x)= \int f(y)P_t(x,dy)\in C_b^\infty\) for \(f\in C_b^\infty\). The author gives a formula of expanding the functional of solutions for (1) into a finite Itô-Wiener chaos with explicit residual. Denote \(\nabla_\sigma f(x)= \sigma(x) (d/dx)f(x)\) and \(P_t'(x,dy) =(d/dx)P_t(x,dy)\), and \(J_n (f_n)\) is the multiple Itô-Wiener integral of order \(n\) [cf. \textit{K. Itô}, J. Math. Soc. Japan 3, 157-169 (1951; Zbl 0044.12202)]. The following is one of the main theorems: For any \(f\in C_b^\infty\), with \(t=1\), \[ f(X_1) =\mathbb{E}_f(X_1)+ \sum^n_{k=1} J_n(f_n) +J_{n+1} (g_{n+1}) \tag{2} \] where \[ \begin{multlined} g_n(s_1, \dots, s_n)= \nabla_\sigma P_{s_2 -s_1} \cdots \nabla_\sigma P_{1-s_n} f(X_{s_1}) \\ \text{and } f_n= \mathbb{E} g_n (s_1, \dots, s_n)=P_{s_1} \nabla_\sigma P_{s_2-s_1} \cdots \nabla_\sigma P_{1-s_n} f(\xi). \end{multlined} \] Note that \(J_{n+1} (g_{n+1})\) is orthogonal to all Itô-Wiener chaoses of order less than or equal to \(n\). As applications of the above, by employing (2) the author derives FKG type inequality, variance inequality for diffusions and correlation inequality for Gaussian measure. A simple proof for Houdré-Kagan's inequality is also given; see \textit{C. Houdré} and \textit{A. Kagan} [J. Theor. Probab. 8, No. 1, 23-30 (1995; Zbl 0815.60018)].

Recommendations

Cited in

(24)

This page was built for publication: Itô-Wiener chaos expansion with exact residual and correlation, variance inequalities

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1374847)