Existence and continuity of occupation densities of stochastic integral processes (Q686784)

Generalizing his earlier papers the author studies conditions for the existence of an occupation density for Skorokhod integral processes in the second Wiener chaos replacing the former approach via Fourier analysis by a probabilistic one: Assuming that \(f\in L^ 2([0,1]^ 2)\) is of the form \(f=\sum^ \infty_{i=1}a_ ih_ i\otimes h_ i\), \((h_ i)\) ONB in \(L^ 2([0,1])\), \(\sum^ \infty_{i=1}| a_ i|^ \alpha<\infty\), for some \(0<\alpha<1/3\), the author studies the Skorokhod integral process \(U_ t=\int^ t_ 0\int^ 1_ 0f(s,r)dW_ sdW_ r\). Putting \[ X_ i=a_ i\int^ 1_ 0h_ i(s)dW_ s\qquad\text{and }\quad S_ t(x)=\sum^ \infty_{i=1}\left(x_ i\int^ t_ 0h_ i(s)dW_ s-a_ i \int^ t_ 0h^ 2_ i(s)ds\right),\;x\in\ell_ 2, \] he has \(U_ t=S_ t((X_ i)_{i\geq 1})\). In a first step the author uses Tanaka's formula to compute the occupation density \(L_ t(x)\) of \(S(x)\); in a second step, applying Kano's and Pisier's generalization of Kolmogorov's continuity criterion he proves a.s. continuity of \(L(x)\) in appropriate bounded subsets of \(\ell_ 2\). This allows the author, finally, to resubstitute \((X_ i(\omega))_{i\geq 1}\) in \(L_ t(x)\) to obtain the occupation density for the process \(U\). Meanwhile, in recent papers with D. Nualart, the author has studied rather general Skorokhod integral processes via an iterated integration by parts formula, cf. \textit{P. Imkeller} and \textit{D. Nualart} [Potential Anal. 2, No. 2, 137-155 (1993; Zbl 0777.60043)].