Ergodic properties of locally stationary processes. (English) Zbl 0612.60035

Let x(t) be a locally stationary process with \(Ex(t)=0\) and the covariance function R(s,t) satisfying the condition: \(R(s,t)=R_ 1(s+t)R_ 2(s-t)\). It is assumed that x(t) is harmonizable, i.e., can be expressed as a stochastic integral: \[ x(t)=\int^{+\infty}_{- \infty}\exp (it\lambda)dX(\lambda) \] where X(\(\lambda)\) is a second order stochastic process with a bounded covariance function. The paper starts with a lemma demonstrating that stochastic integrals of x(t) can be expressed as integrals of X(\(\lambda)\): \[ \int^{T}_{- T}x(t)dt=\int^{+\infty}_{-\infty}\{\int^{T}_{-T}\exp (it\lambda)dt\}dX(\lambda). \] Three ergodic theorems follow, for various sets of assumptions on the covariance function R(s,t).
Reviewer: L.Podkaminer


60G10 Stationary stochastic processes
60H05 Stochastic integrals
60F20 Zero-one laws
60E10 Characteristic functions; other transforms
Full Text: EuDML


[1] J. Michálek: Spectral decomposition of locally stationary random processes. Kybernetika 22 (1986), 3, 244-255. · Zbl 0601.60036
[2] M. Loève: Probability Theory. D. van Nostrand, Toronto-New York-London 1955. · Zbl 0066.10903
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.