Upper and lower estimates for integer complexity
Abstract
Let stand for the integer complexity of the number , i.e. for the least number of 's needed to write using arbitrary many additions, multiplications, and parentheses. The two-sided inequality for all is well known and reveals the logarithmic behaviour of the complexity function . While the lower bound is attained infinitely many times at powers of , the best upper estimate is still unknown, although there are some improvements of the trivial bound . Besides, for typical numbers, i.e. for almost all numbers , the better inequality holds, where, importantly, . We show that in fact as , which, in particular, yields that . We also obtain the first nontrivial lower bound for almost all numbers .
Cite
@article{arxiv.2603.20876,
title = {Upper and lower estimates for integer complexity},
author = {Sergei Konyagin and Kristina Oganesyan},
journal= {arXiv preprint arXiv:2603.20876},
year = {2026}
}
Comments
14 pages. The value of $C_{avg}$ is changed and several inaccuracies are corrected