Exponential tightness and projective systems in large deviation theory
Let \(E\) be a Hausdorff topological space with a \(\sigma\)-field \({\mathcal E}\) containing the class of compact sets and a basis for the topology of \(E\). Moreover, let \(\{E_\alpha, p^\beta_\alpha\}\), \(\alpha,\beta\in A\), be a projective system of Hausdorff topological spaces indexed by a directed set \(A\) and let \(p_\alpha: E\to E_\alpha\), \(p_\alpha= p^\beta_\alpha\circ p_\beta\), be a separating the points of \(E\) family of continuous surjective maps which are \({\mathcal E}\)-measurable when \(E_\alpha\) is endowed with the Borel \(\sigma\)-field. Assuming that \(\{\mu_d\}\), \(d\in D\) a directed set, is a net of p.m.'s on \({\mathcal E}\) such that (i) for each \(\alpha\in A\), \(\{\mu_d\circ p^{-1}_\alpha\}\), \(d\in D\), satisfies the large deviation principle with normalizing constants \(\{r_d\}\), \(\lim_d r_d= \infty\), and rate function \(I_\alpha: E_\alpha\to [0, \infty]\), and (ii) \(\{\mu_d\}\) is exponentially tight, the author proves that \(\{\mu_d\}\) then satisfies the large deviation principle with the same normalizing constants and rate function \(I(x)= \sup_\alpha I_\alpha(p_\alpha(x))\), \(x\in E\), where \(\{x\in E: I(x)\leq a\}\) is compact for every \(a\geq 0\). Next, on the basis of such a general theorem, large deviation principles for normalized sums of i.i.d. r.v.'s in a separable Banach space and a sequence of \(\mathbb{R}^d\)-valued Lévy processes are derived. The results obtained in this way encompass some theorems proved earlier by \textit{M. D. Donsker} and \textit{S. R. S. Varadhan} [Commun. Pure Appl. Math. 29, 389-461 (1976; Zbl 0348.60032)], \textit{R. R. Bahadur} and \textit{S. L. Zabell} [Ann. Probab. 7, 587-621 (1979; Zbl 0424.60028)], \textit{R. Azencott} [in: École d'été de probabilités de Saint-Flour VIII--1978. Lect. Notes Math. 774, 1-176 (1980; Zbl 0435.60028)], the author [in: Probability in Banach sapces V. Lect. Notes Math. 1153, 1-14 (1985; Zbl 0571.60038)], \textit{D. A. Dawson} and \textit{J. Gärtner} [Stochastics 20, 247-308 (1987; Zbl 0613.60021)], \textit{A. Pukhalskij} [in: New trends in probability and statistics. Vol. 1, 198-218 (1991; Zbl 0767.60024)] and the author [Stochastic Processes Appl. 51, No. 1, 75-115 (1994; Zbl 0805.60020)]. For the entire collection see [Zbl 0861.00032].
- Trajectories of exchangeable sequences: Large and moderate deviations results
- Large and moderate deviations for the empirical measures of an exchangeable sequence
- Large and moderate deviations for infinite-dimensional autoregressive processes.
- Large deviations for Poisson random measures and processes with independent increments
- Large deviation principle for stochastic integrals and stochastic differential equations driven by infinite-dimensional semimartingales
- Moderate deviations for Markov chains with atom.
- Functional approach of large deviations in general spaces
- Large deviation for diffusions and Hamilton-Jacobi equation in Hilbert spaces
- Large deviations of semimartingales: A maxingale problem approach i. limits as solutions to a maxingale problem
- Large deviation principle for the greedy exploration algorithm over Erdős-Rényi graphs
- A large deviation principle for the Schramm–Loewner evolution in the uniform topology
- Large deviations for Markov processes with switching and homogenisation via Hamilton-Jacobi-Bellman equations
- Large deviations for the greedy exploration process on configuration models
- Moderate deviations for some point measures in geometric probability
This page was built for publication: Exponential tightness and projective systems in large deviation theory
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4355981)
