Lower bounds for densities of uniformly elliptic random variables on Wiener space (Q1400843): Difference between revisions

The first aim of this work is to supply some general criteria insuring that a Wiener functional \(F\) has a density admitting some Gaussian-like lower bound, thereby generalizing previous results established by Kusuoka and Stroock for some special diffusions. Some variants of such criteria are given, based on the assumption that \(F\) has along any partition of time conditional densities possessing sufficient lower bounds. Such statements are meant to possibly apply even in the absense of Markov property and of any special Lie algebra structure. As an application, it is shown that this sort of criteria works in the case of the solution to a uniformly elliptic stochastic heat equation with Neumann conditions, getting indeed there for the density \(p_t(x,y)\) the explicit lower bound \(\varepsilon t^{-1/4}\exp[-\|y- g_t(x)\|^2/\varepsilon\sqrt{t}]\).