Product representation of perfect cubes

TL;DR

This paper investigates the maximum size of subsets of [n] avoiding solutions to a specific product equation involving perfect cubes, providing bounds for various parameters and refuting a long-standing conjecture.

Contribution

It extends the study of product-free sets to the case of perfect cubes, offering new bounds and disproving an existing conjecture in the field.

Findings

Established bounds for F_{k,3} for k=2,3,4,6,9

Refuted an 18-year-old conjecture of Verstra"ete

Introduced a related function f_{k,d} with relaxed conditions

Abstract

Let $F_{k,d}(n)$ be the maximal size of a set ${A}\subseteq [n]$ such that the equation \[a_1a_2\dots a_k=x^d, \; a_1<a_2<\ldots<a_k\] has no solution with $a_1,a_2,\ldots,a_k\in {A}$ and integer $x$. Erd\H{o}s, S\'ark\"ozy and T. S\'os studied $F_{k,2}$, and gave bounds when $k=2,3,4,6$ and also in the general case. We study the problem for $d=3$, and provide bounds for $k=2,3,4,6$ and $9$, furthermore, in the general case, as well. In particular, we refute an 18 years old conjecture of Verstra\"ete. We also introduce another function $f_{k,d}$ closely related to $F_{k,d}$: While the original problem requires $a_1, \ldots , a_k$ to all be distinct, we can relax this and only require that the multiset of the $a_i$'s cannot be partitioned into $d$-tuples where each $d$-tuple consists of $d$ copies of the same number.