Efficient Monte Carlo for High Excursions of Gaussian Random Fields (with R. Adler and J. C. Liu).

Summary:

This paper concentrates on sampling of continuous Gaussian random fields conditional on large high excursions in polynomial time (polynomial in the excursion level). The algorithms in the paper assume only modest continuity assumptions both in the drift and the covariance. Optimal (in the sense of obtaining basically bounded complexity in the level) algorithms for smooth fields are discussed. The paper is meaningful because, despite the extensive work on asymptotics for the maximum of Gaussian random fields, the algorithms enable fast computation of functional of the whole fields given high excursions.

Bibtex:

@Article{AlderBlanchetLiu10,

author = {Adler, R. and Blanchet, J. and Liu, J. C.},

title = {Efficient Monte Carlo for High Excursions of Gaussian Random Fields},

journal = {Pre-print},

year = {2010},

volume = {},

pages = {}

}