Some Optimal Conditions for the ASCLT

Abstract

Let $X_1,X_2,\dots$ be independent random variables with $E{X}_k=0$ and ${\sigma }_k^2:=E{X}_k^2<\infty$ $(k\ge 1)$. Set $S_k=X_1+\cdots +X_k$ and assume that $s_k^2:=E{S}_k^2\to \infty$. We prove that under the Kolmogorov condition $\begin{array}{c}|X_n|\le L_n,L_n=o(s_n/{(loglog{s}_n)}^{1/2})\end{array}$we have $\begin{array}{c}\frac{1}{log{s}_n^2}\sum _{k=1}^n\frac{{\sigma }_{k+1}^2}{s_k^2}f\left(\frac{S_k}{s_k}\right)\to \frac{1}{\sqrt{2\pi }}{\int }_{R}f\left(x\right)e^{-x^2/2}\text{d}xa.s.\end{array}$for any almost everywhere continuous function $f:R\to R$ satisfying $\left|f\left(x\right)\right|\le e^{\gamma x^2}$, $\gamma <1/2$. We also show that replacing the o in (1) by O, relation (2) becomes generally false. Finally, in the case when (1) is not assumed, we give an optimal condition for (2) in terms of the remainder term in the Wiener approximation of the partial sum process $\{S_n,n\ge 1\}$ by a Wiener process.

Keywords: Almost sure central limit theorem; Sums of independent random variables; Weighted averages.