Precise asymptotics in Chung's law of the iterated logarithm (Q943505): Difference between revisions

Consider an i.i.d. sequence \(X,X_1,X_2,\dots\) of random variables with mean zero, positive variance \(\sigma^2,\) and \(E\,X^2I\{| X| \geq t\}=o((\log\log t)^{-1})\) as \(t\to\infty.\) Set \(S_n=X_1 +\dots+X_n\) and \(M_n=\max_{k\leq n}| S_k| ,\) \(n\geq 1.\) In this paper, a number of precise asymptotics is proved related to Chung's law of the iterated logarithm. For example, if \(\{a_n\}\) is a real sequence such that \(a_n=o(1/\log\log n),\) then, for any \(a>-1\) and \(b>-1,\) \[ \begin{multlined} \lim_{\varepsilon\nearrow 1/\sqrt{1+a}}\left(\frac{1}{\sqrt{1+a}}-\varepsilon\right)^{b+1} \sum_{n=1}^\infty \frac{(\log n)^a (\log\log n)^b}{n} P\left\{M_n\leq \sqrt{\frac{\sigma^2 \pi^2 n}{8\log\log n}}\,(\varepsilon +a_n)\right\} \\ =\frac{4}{\pi}\left(\frac{1}{2(1+a)^{3/2}}\right)^{b+1}\Gamma (b+1),\end{multlined} \] where \(\Gamma\) denotes the gamma function. The results are first obtained under normal distribution, after which it is be shown that, via truncation and approximation, the same asymptotics can be proved under a general distribution, too. Moreover, the author provides necessary and sufficient conditions for his results to hold.