On large deviations of coupled diffusions with time scale separation

DOI10.1214/15-AOP1043MaRDI QIDQ317501zbMATH OpenOpenAlexFDO

Publication date 30 September 2016

Published in The Annals of Probability (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1306.5446, https://projecteuclid.org/euclid.aop/1470139161

averaging coupled diffusions empirical process large deviation principle time scale separation

Large deviations (60F10) Functional limit theorems; invariance principles (60F17) Random measures (60G57) Stochastic ordinary differential equations (aspects of stochastic analysis) (60H10) Diffusion processes (60J60)

Abstract: We consider two Ito equations that evolve on different time scales. The equations are fully coupled in the sense that all coefficients may depend on both the "slow" and the "fast" processes and the diffusion terms may be correlated. The diffusion term in the "slow" process is small. A large deviation principle is obtained for the joint distribution of the slow process and of the empirical measure of the fast process. By projecting on the slow and fast variables, we arrive at new results on large deviations in the averaging framework and on large deviations of the empirical measures of ergodic diffusions, respectively. The proof of the main result relies on the property that exponential tightness implies large deviation relative compactness. The identification of the large deviation rate function is accomplished by analysing the large deviation limit of an exponential martingale.

Review text

Consider the coupled diffusions given by NEWLINE\[NEWLINE\begin{aligned} dX^\varepsilon_ t &= A(X^\varepsilon_t, x^\varepsilon_t)\,dt+ \sqrt{\varepsilon} B(X^\varepsilon_t, x^\varepsilon_t)\,dW^\varepsilon_t,\\ dx^\varepsilon_t &= {1\over \varepsilon} a(X^\varepsilon_t,x^\varepsilon_t)\,dt+ {1\over\sqrt{\varepsilon}} b(X^\varepsilon_t, x^\varepsilon_t)\,dW^\varepsilon_t,\end{aligned}NEWLINE\]NEWLINE where \(\varepsilon> 0\) is a small parameter, \(A(u,x)\) an \(n\)-vector with \(u\in\mathbb{R}^n\) and \(x\in \mathbb{R}^1\), \(B(u,x)\) an \(n\times k\)-matrix, \(a(u,x)\) an \(\ell\)-vector, \(b(u,x)\) an \(\ell\times k\)-matrix, and \(W^\varepsilon\) is an \(\mathbb{R}^k\)-valued standard Wiener process. The processes \(X^\varepsilon\) and \(x^\varepsilon\) are evolving on different time scales, the time for \(x^\varepsilon\) being accelerated by a factor \(1/\varepsilon\). Denote by \(\mu^\varepsilon\) the empirical process associated with \(x^\varepsilon\).NEWLINENEWLINE The author obtains a large deviation principle (LDP) for the distribution of \((X^\varepsilon,\mu^\varepsilon)\) and the large deviation rate function. Projections on the first and the second component lead to LDPs for \(X^\varepsilon\) and \(\mu^\varepsilon\), respectively. This result generalizes that of \textit{R. Liptser} [Probab. Theory Relat. Fields 106, No. 1, 71--104 (1996; Zbl 0855.60030)], who treated the case of one-dimensional \(X^\varepsilon\) and \(x^\varepsilon\) with \(a\) and \(b\) not depending on \(u\) and the Wiener processes being independent.

Recommendations

Cites work

Cited in

(37)

This page was built for publication: On large deviations of coupled diffusions with time scale separation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q317501)