Lower bounds for the density of locally elliptic Itô processes (Q874742)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Lower bounds for the density of locally elliptic Itô processes |
scientific article |
Statements
Lower bounds for the density of locally elliptic Itô processes (English)
0 references
10 April 2007
0 references
The main interest of the author of this paper is to give an explicit lower bound for the density \(p_T(x_0,y)\) of the law of \(X_T\), solution of the SDE on \({\mathbb R}^q\) \[ dX_t=\sigma(X_t) dB_t + b(X_t)dt,\quad X_0=x_0. \] A subject which was intensively discussed during the 1980s. The novelty is that the author assumes only the ellipticity along a differentiable curve joining \(x_0\) and \(y\) (e.g., a tubular neighborhood), apart from the non-boundedness hypothesis on the diffusion coefficient. More precisely, let \(\lambda^*(x)\geq \sigma\sigma^*(x)\geq \lambda_*(x)\), consider the class \(\Phi(x_0,y)\) of functions \(\varphi\in L^2([0,T], {\mathbb R}^d)\) satisfying the following conditions \[ x_0^\varphi=x_0, x_T^\varphi=y,\quad {\lambda_*\over\lambda^*}(x_t^\varphi)\geq {1\over\mu},\;\lambda_*(x_t^\varphi)\geq 1/\chi, \] where \(\mu, \chi\) are given and \[ dx_t^\varphi=\sum_{j=1}^d \sigma_j(x_t^\varphi)\varphi_t^j\, dt,\quad x_0^\varphi=x_0, \] and define the control distance \[ d(x_0,y)=\inf\biggl\{\sqrt{\int_0^T | \varphi_t| ^2\,dt}; \varphi\in\Phi(x_0,y)\biggr\}. \] Then a Gaussian type lower bound for \(p_T(x_0,y)\) is obtained using \(d(x_0,y), \mu\) and \(\chi\). Finally, the author develops the conditional version of the Malliavin calculus on one hand and uses the methodology put forward on Wiener functionals by Kohatsu-Higa on other hand. As byproducts, the author obtains general results about the so-called elliptic Itô processes.
0 references
Gaussian lower bound
0 references
density
0 references
Malliavin calculus
0 references
control distance
0 references
local ellipticity
0 references