English

Upper and lower estimates for integer complexity

Number Theory 2026-05-01 v2 Combinatorics

Abstract

Let n\|n\| stand for the integer complexity of the number nn, i.e. for the least number of 11's needed to write nn using arbitrary many additions, multiplications, and parentheses. The two-sided inequality 3log3nn3log2n3\log_3 n\leq\|n\|\leq 3\log_2 n for all nn is well known and reveals the logarithmic behaviour of the complexity function n\|n\|. While the lower bound 3log3n3\log_3 n is attained infinitely many times at powers of 33, the best upper estimate is still unknown, although there are some improvements of the trivial bound 3log2n3\log_2 n. Besides, for ``typical"" numbers, i.e. for almost all numbers nn, the better inequality nCavglogn\|n\|\leq C_{avg}\log n holds, where, importantly, Cavg3.236<supnnlognC_{avg}\approx 3.236<\sup_{n} \frac{\|n\|}{\log n}. We show that in fact nCavglogn+o(logn)\|n\|\leq C_{avg}\log n+o(\log n) as nn\to\infty, which, in particular, yields that lim supnnlognCavg\limsup\limits_{n\to\infty}\frac{\|n\|}{\log n}\leq C_{avg}. We also obtain the first nontrivial lower bound n3.06log3n\|n\|\geq 3.06\log_3 n for almost all numbers nn.

Keywords

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