Lévy area for Gaussian processes: a double Wiener-Itô integral approach (Q553091): Difference between revisions

Given two independent processes \(X_1=(X_1(t))_{0\leq t\leq 1}\) and \(X_2=(X_2(t))_{0\leq t\leq 1}\), let \(A\) be the Lévy area generated by them, i.e., \(A\) is the area included by the curve \(x=X_1(t)\), \(y=X_2(t)\) and by its cord. In the case that \(X_1\) and \(X_2\) are Wiener processes, the random variable \(A\) was completely described by Paul Lévy in 1951. The aim of the present paper is to find a description of \(A\) for more general centered Gaussian continuous processes \(X_1\) and \(X_2\). Under some assumptions on the variation of \(X_1\) and \(X_2\), the authors represent \(A\) as a double Wiener-Itô integral with respect to an isonormal Gaussian process induced by \(X_1\) and \(X_2\).