Variance of statistics defined on random permutations
摘要
The present work improves the authors’ result [A. Karbonskis and E. Manstavičius, Variance of a strongly additive function defined on random permutations, Lith. Math. J., 64(3):302–314, 2024] on the variance of a strongly additive function defined on the symmetric group endowed with the uniform probability measure. Previously, we established an upper estimate containing the optimal asymptotic constant; however, the remainder term bound, when compared to that obtained by Kubilius [J. Kubilius, Improved estimate of the second central moment for additive arithmetic functions, Lith. Math. J., 25(3):250–254,1985] in number theory, remained unsatisfactory. Inspired by his result, we now reach an analogous level by applying a novel methodology. In particular, we take advantage of the power method in the eigenvalue localization algorithm.