Convergence rates of limit distribution of maxima of lognormal samples (Q448264): Difference between revisions

It is known that linearly normalized partial maxima of independent and identically distributed (iid) random variables (rvs) from the lognormal distribution converge weakly to a rv having the Gumbel distribution \(\Lambda.\) In this paper the authors prove the following result: If \(\{X_n, n \geq 1\}\) is a sequence of iid rvs with common lognormal distribution \(F\) and if the norming constants \(a_n\) and \(b_n\) are given by the solution of the equations \[ 2\pi (\log b_n)^2exp(\log b_n)^2=n^2,\;a_n=b_n/\log{b_n}, \] then there exist absolute constants \(0< C_1 < C_2\) such that \[ C_1/\sqrt{\log n} < \sup_{x \in R}\mid F^n(a_n x + b_n) - \Lambda(x)\mid <C_2/\sqrt{\log n}, \] for \(n \geq 2\). And for the norming constants \[ \alpha_n=exp{\sqrt{2 \log n}/\sqrt{2 \log n}} \] and \[ \beta_n = \exp \sqrt{2 \log n}\left( 1 - \frac{\log 4\pi + \log \log n}{2 \sqrt{2 \log n}}\right), \] \[ F^n(\alpha_n x + \beta_n) - \Lambda(x)\sim -\frac{e^{-x}\exp e^{-x}}{8} \frac{(\log \log n)^2}{\sqrt{2 \log n}} \] for large \(n\).