Random fields and the geometry of Wiener space (Q359687): Difference between revisions

The main object of study in this paper is an (extended) class of smooth random fields which can be expressed as smooth Wiener functionals. In this context, the authors extend the so-called Gaussian Minkowski functionals (GMFs), which are finite-dimensional Gaussian geometric functionals arising in the so-called Gaussian tube formula, to infinite dimensions. In order to achieve this, the authors first modify the definition of GMFs in finite dimensions in a suitable way which can be more easily extended to the infinite-dimensional case. The extension is then described in detail and Theorem 5.1 (p.~2745) contains one of the key results of the paper. Theorem 5.1. Let \(F\in D_{2+\delta}^{\infty-}(X;\mathbb{R})\) be an \(H\)-convex Wiener functional such that it satisfies all the regularity conditions of Theorem 4.2, and \(A_{u}=F^{-1}(-\infty,u]\), then \[ \mu(\mathrm{Tube}(A_{u},\rho)) = \mathcal{M}_0^{\mu}( A_{u})+ \sum_{j=1}^{\infty}\frac{\rho^j}{j!}\mathcal{M}_j^{\mu}( A_{u}), \] where \(\mathcal{M}_j^{\mu}( A_{u})\) are the infinite-dimensional versions of Gaussian Minkowski functionals and, as usual, \(\mathcal{M}_0^{\mu}( A_{u})=\mu(A_{u})\). The paper concludes with some applications of the results to random fields which can be expressed in terms of stochastic integrals driven by a Gaussian field.