The almost sure central limit theorems in the joint version for the maxima and sums of certain stationary Gaussian sequences (Q2479327): Difference between revisions

Let \(x_{1},x_{2},\dots ,x_{n}\) be a standardized stationary Gaussian sequence, \[ M_{n}=\max(x_{1},\dots,x_{n}),S_{n}=x_{1}+..+x_{n},\sigma _{n}=\sqrt{ V_{ar}S_{n}}. \] The covariance function \(r(t)=\text{Cov}(x_{1},x_{1+t})=L(t)t^{-\alpha },t=1,2,\dots ,\alpha >0,\) where \(L(t)\) is a positive slowly varying function at infinity. The normal distribution function is denoted by \(\Phi\). The following statements are proved (i) If the numerical sequence \(\{u_{n}\}\) satisfies the relation \(\underset{N\rightarrow \infty }{\lim}n\left( 1-\Phi \left( u_{n}\right) \right) =\tau\), for some, \(0\leqslant \tau \leqslant \infty \), then \[ \lim_{N\rightarrow \infty }\;\frac{1}{\log N}\overset{N}{\underset{ n=1}{\sum }}\frac{1}{n}\mathbf{1}\left\{ M_{n}\leqslant u_{n},S_{n}/\sigma _{n}\leqslant y\right\} =e^{-\tau }\Phi (y) \] for any \( y\leqslant R.\) \noindent (ii) If for \[ n\geqslant 3,a_{n}=(2\log n)^{\frac{1}{2}},b_{n}=(2\log n)^{\frac{1}{2}}-\tfrac{1}{2}\left( 2\log n\right) ^{-\frac{1}{2}}\left( \log \log n+\log 4\pi \right) , \] then as a direct conclusion it follows that \[ {N\rightarrow \infty }{\lim}\frac{1}{\log N}\overset{N}{\underset{ n=1}{\sum }}\frac{1}{n}\mathbf{1}\left \{ a_{n}\left( M_{n}-b_{n}\right) \leqslant x,\frac{S_{n}}{\sigma _{n}}\leqslant y\right \} =\exp \left( -e^{-x}\right) \Phi \left( y\right) \] a.s. for all \(x,y\in R\).