Small time and semiclassical asymptotics for stochastic heat equations driven by Lévy noise (Q4804690)

Let \(\xi\) be an \(\mathbb R^{d}\)-valued Lévy process containing no Brownian component, let \(\nu\) be its Lévy measure. Assume that \(a, V:\mathbb R^{d}\to\mathbb R\), \(c:\mathbb R^{d}\to\mathbb R^{d}\) are measurable functions, \(c(x)y\geq 0\) for all \(x\in \mathbb R^{d}\), \(y\in\text{supp }\nu\), \(h>0\). A stochastic heat equation \[ h d\psi(t,x) = - \Bigl(-{h^2\over 2}\Delta + V(x) + ha(x)\Bigr)\psi(t,x) dt + h\psi(t-,x)c(x) d\xi_{t} \tag{1} \] is studied. Under smoothness and boundedness hypotheses on the functions \(V\), \(a\) and \(c\) small time and/or small \(h\) asymptotics for the Green function (the fundamental solution) of (1) and for solutions to the Cauchy problem for (1) are found by applying the WKB method. For example, the following large deviation principle is established: if \(\psi_{G}\) is the fundamental solution of (1), then \[ \lim_{h\to 0} h\log \psi_{G}(t,t_0,x,x_0) = -S(t,t_0,x,x_0) \] and \[ \lim_{t \to t_0} (t-t_0)\log\psi_{G}(t,t_0,x,x_0) = - (2h)^{-1} |x-x_0|^2, \] where \(S(\cdot,t_0,\cdot,x_0)\) is a suitable solution to the stochastic Hamilton-Jacobi equation \[ dS + \frac 12 (\partial S/\partial x)^2 dt - V dt - ha dt + hc d\xi_{t}=0 \] on \(\mathopen]t_0,T \mathclose[\times \mathbb R^{d}\).