Exponential decay or correlation for the stochastic process associated to the entropy penalized method

Abstract: In this paper we present an upper bound for the decay of correlation for the stationary stochastic process associated with the Entropy Penalized Method. Let

L (x, v) : T t^{n} i m e s R r^{n} o R r

be a Lagrangian of the form L(x,v) = {1/2}|v|^2 - U(x) + < P, v>. For each value of

e p s i l o n

and

h

, consider the operator Gg[phi](x):= -epsilon h {ln}[int_{ e^N} e ^{-frac{hL(x,v)+phi(x+hv)}{epsilon h}}dv], as well as the reversed operator �ar Gg[phi](x):= -epsilon h {ln}[int_{ e^N} e^{-frac{hL(x+hv,-v)+phi(x+hv)}{epsilon h}}dv], both acting on continuous functions

p h i : T t^{n} o R r

. Denote by

p h i_{e p s i l o n, h}

the solution of

G g [p h i_{e p s i l o n, h}] = p h i_{e p s i l o n, h} + l a m b d a_{e p s i l o n, h}

, and by

the solution of

. In order to analyze the decay of correlation for this process we show that the operator

c a l L (p h i) (x) = i n t e^{- f r a c h L (x, v) e p s i l o n} p h i (x + h v) d v,

has a maximal eigenvalue isolated from the rest of the spectrum.

Review text

Let \(T^n\) be the \(n\)-dimensional torus. The authors consider the Lagrangian \(L(x,v)\: T^n\times \mathbb{R}^N\to \mathbb{R}\) of the form \[ L(x,v) = \frac12| v| ^2 - U(x) + \{ P,v\}, \] where \(U\in C^1(T^n)\) and \(P\in \mathbb{R}^n\) is constant. The paper deals with the discrete time Aubry--Mather problem and the entropy penalized Mather method which provides a way to obtain approximations by continuous densities of the Aubry--Mather measure.

Recommendations

Cited in

(1)

The mather measure and a large deviation principle for the entropy penalized method

This page was built for publication: Exponential decay or correlation for the stochastic process associated to the entropy penalized method

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2390109)