Some results on stochastic convolutions arising in Volterra equations perturbed by noise (Q1357119)

Let \(H\) be a separable Hilbert space and let \(\{e_k\}\) be a complete orthonormal system in \(H\). The authors consider a stochastic version of Volterra equation in \(H\) of the general form: \[ X(t)=\int^t_0 a(t-s)AX(s)ds+x+W(t),\qquad x\in H,\tag{*} \] where \(W(t)\) is a cylindrical Wiener process defined in a stochastic basis \((\Omega,{\mathcal F},{\mathcal F}_t,\mathbb{P})\) such that \[ \langle W(t),h\rangle=\sum^\infty_{k=1} \langle h,e_k \rangle \beta_k(t),\qquad h\in H, \] with a sequence \(\{\beta_k\}\) of real-valued mutually independent Wiener processes. According to the semigroup approach [cf. \textit{G. Da Prato} and \textit{J. Zabczyk}, ``Stochastic equations in infinite dimensions'' (1992; Zbl 0761.60052)], the mild solution of (*) is an \(\{{\mathcal F}_t\}\)-adapted process \(X(t)\), \(t\geq 0\), such that \[ X(t)=S(t)x+W_{A,a}(t),\text{ with }W_{A,a}(t)=\int^t_0 S(t-s)dW(s). \] We call \(W_{A,a}\) the stochastic convolution. The authors are interested in deriving regularity properties of \(W_{A,a}\). Assume that \(a(t)\in L^1_{\text{loc}}(0,+\infty)\) is completely positive and \(A\) is a linear self-adjoint negative operator in \(H\) such that \(-\text{Tr}(A^{-1})=\sum^\infty_{k=1} (1/\mu_k) <+\infty\), where \(Ae_k=-\mu_k e_k\), \(k\in \mathbb{N}\), for some positive number \(\mu_k\). Let \(H=L^2({\mathcal O})\) with a bounded open subset \(\mathcal O\) of \(\mathbb{R}^n\). Under some additional technical conditions, the authors prove that the trajectories of \(W_{A,a}(t)(\xi)\) \((\xi\in {\mathcal O})\) are almost surely \(\alpha\)-Hölder continuous in \(t\), \(\xi\) for any \(\alpha\in (0,1/4)\).