The limit theorems for maxima of stationary Gaussian processes with random index

From MaRDI portal
Publication:2453856





Let \(\{ X( t) :t\geq0\} \) be a standard (zero-mean, unit-variance) stationary Gaussian process with continuous sample paths and correlation function \(r( \cdot) \) satisfying the following condition \(r( t) =1-C| t| ^{\alpha}+o( | t| ^{\alpha}) \) as \(t\rightarrow0\) and \(r( t) <1\) for \(t>0\) for some positive constant \(C\) and \(\alpha\in( 0,2] \). The author proves limit theorems for the maxima \[ M( \mathcal{T}_{T}) =\max\{ X( t) : t\in[ 0,\mathcal{T}_{T}] \} \text{ as }T\rightarrow\infty\text{,} \] where the random index \(\mathcal{T}_{T}\) is a positive random process. Two cases are considered: (1) \(\mathcal{T}_{T}\) is independent of the basic process \(\{ X( t) : t\geq0\} \), \(\mathcal{T}_{T}/T\) converges weakly to some non-degenerate distribution, and \(\lim_{T\rightarrow\infty}r( T) \ln T=r\in[ 0,\infty] \); (2) the interrelation of \(\{ X( t): t\geq0\} \) and \(\mathcal{T}_{T}\) is not restricted, \(\mathcal{T}_{T}/T\rightarrow\mathcal{T}\) in probability, where \(\mathcal{T}\) is a positive random variable, and \(\lim_{T\rightarrow\infty}r( T) ( \ln T) ^{\frac{2+\alpha}{2\alpha}}=0\).



Cites work









This page was built for publication: The limit theorems for maxima of stationary Gaussian processes with random index

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2453856)