Extremes of multidimensional stationary Gaussian random fields

Abstract: Let

X (m a t h b f t) : m a t h b f t = (t_{1}, t_{2}, l d o t s, t_{d}) i n [0, i n f t y)^{d}

be a centered stationary Gaussian field with almost surely continuous sample paths, unit variance and correlation function

r

satisfying conditions

r (m a t h b f t) < 1

for every

m a t h b f t e q m a t h b f 0

and

r (m a t h b f t) = 1 - s u m_{i = 1}^{d} | t_{i} |^{a l p h a_{i}} + o (s u m_{i = 1}^{d} | t_{i} |^{a l p h a_{i}})

, as

m a t h b f t o m a t h b f 0

, with constants

a l p h a_{1}, a l p h a_{2}, l d o t s, a l p h a_{d} i n (0, 2]

. The main result of this contribution is the description of the asymptotic behaviour of

P (s u p X (m a t h b f t) : m a t h b f t i n m a t h c a l J_{m a t h b f m}^{m a t h b f x} l e q s l a n t u)

, as

u o i n f t y

, for some Jordan-measurable sets

m a t h c a l J_{m a t h b f m}^{m a t h b f x}

of volume proportional to

P (s u p X (m a t h b f t) : m a t h b f t i n [0, 1]^{d} > u)^{- 1} (1 + o (1))

.

Recommendations

Cites work

Cited in

(16)

This page was built for publication: Extremes of multidimensional stationary Gaussian random fields