A stochastic process generated by the Lévy Laplacian (Q5937979)

Let \(\Phi\) be a generalized white noise functional. The \(S\)-transform is defined as \(S[\Phi](\xi)=\exp(-{{1}\over{2}}\langle \xi,\xi\rangle)\) \(\langle \Phi(\cdot),\exp\langle \cdot,\xi\rangle\rangle\), where \(\exp\langle \cdot,\xi\rangle\) is a testing functional generated by a function \(\xi\) on the finite interval \(T\). The Lévy Laplacian \(\Delta_L\) is defined as \[ \Delta_L\Phi=S^{-1}\widetilde{\Delta}_L S[\Phi] \quad\text{and}\quad \widetilde{\Delta}_L S[\Phi](\xi)=\lim_{n\rightarrow \infty}{{1}\over{N}} \sum_{n=0}^{N-1}S[\Phi]''(\xi)(\zeta_n,\zeta_n), \] where \(S[\Phi]''\) means the Fréchet derivative of 2nd order and \(\{\zeta_n\}\) is a base of \(L^2(T)\). The author considers the stochastic process generated by these Laplacians and obtains the following relations: \(e^{-t(-\widetilde{\Delta}_L)^\alpha} F(\xi)=E[F(\xi+X_t^\alpha\cdot\eta_T)]\), \(0<\alpha\leq 2\), where \(X_t^\alpha\) is the strict stable process of index \(\alpha\), especially for the Cauchy process \(e^{t\widetilde{\Delta}_L}F(\xi)=E[F(\xi+X_t^1 \cdot \eta_T)]\), and for the Gaussian case, \(e^{-t\widetilde{\Delta}_L^2}F(\xi)=E[F(\xi+B_{2t} \cdot \eta_T)]\), where \(\eta_T\) is a smooth function which is constant \(1/|T|\) on \(T\). Finally, \(e^{-t(-\widetilde{\Delta}_L)^\alpha} \Phi(x) = E[\Phi(x+X_t^\alpha \cdot \eta_T)]\).