Asymptotic for the principal eigenvalue and eigenfunction of a nearly first-order operator with large potential (Q1381574): Difference between revisions
From MaRDI portal
Created a new Item |
Added link to MaRDI item. |
||
links / mardi / name | links / mardi / name | ||
Revision as of 15:44, 31 January 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Asymptotic for the principal eigenvalue and eigenfunction of a nearly first-order operator with large potential |
scientific article |
Statements
Asymptotic for the principal eigenvalue and eigenfunction of a nearly first-order operator with large potential (English)
0 references
18 March 1998
0 references
The authors consider operators, which are the sum of the generator of \(X_t\), a diffusion process in \(\mathbb{R}^n\) with small noise and a potential function \(l\). The process \(X_t\) satisfies a stochastic differential equation of the form \(dX_t= g(X_t)dt +\varepsilon^{1/2} dB_t\), where \(B_t\) is an \(n\)-dimensional Brownian motion and \(g:\mathbb{R}^n \to\mathbb{R}^n\) is such that the process \(X_t\) is ergodic. Therefore, operators of the form \[ G^\varepsilon f= (\varepsilon/2) \Delta f+g \cdot \nabla f+(l/ \varepsilon)f \] are considered. The asymptotic behaviour, as \(\varepsilon\to 0\), of the principal eigenvalue \(\lambda^\varepsilon\) and the corresponding normalized eigenfunction \(\psi^\varepsilon\) of the operator \(G^\varepsilon\) are investigated. The control interpretation of \(\lambda^\varepsilon\) and \(\psi^\varepsilon\) is used. Two applications to robust and risk-sensitive control are given.
0 references
stochastic differential equation
0 references
Brownian motion
0 references
principal eigenvalue
0 references
risk-sensitive control
0 references