Equivalences between Non-trivial Variants of 3LDT and Conv3LDT

Abstract

The popular 3SUM conjecture states that there is no strongly subquadratic time algorithm for checking if a given set of integers contains three distinct elements $x_1, x_2, x_3$ such that $x_1+x_2=x_3$. A closely related problem is to check if a given set of integers contains distinct elements satisfying $x_1+x_2=2x_3$. This can be reduced to 3SUM in almost-linear time, but surprisingly a reverse reduction establishing 3SUM hardness was not known. We provide such a reduction, thus resolving an open question of Erickson. In fact, we consider a more general problem called 3LDT parameterized by integer parameters $\alpha_1, \alpha_2, \alpha_3$ and $t$. In this problem, we need to check if a given set of integers contains distinct elements $x_1, x_2, x_3$ such that $\alpha_1 x_1+\alpha_2 x_2 +\alpha_3 x_3 = t$. We prove that all non-trivial variants of 3LDT over the same universe $[-n^c,n^c]$ for some $c\geq2$ are equivalent under subquadratic reductions. The main technical tool used in our proof is an application of the famous Behrend's construction that partitions a given set of integers into few subsets that avoid a chosen linear equation. We extend our results to Conv3LDT and show that for all $c\geq2$, all non-trivial variants of 3LDT over the universe $[-n^c,n^c]$ and of Conv3LDT over the universe $[-n^{c-1},n^{c-1}]$ are subquadratic-equivalent, so in particular they are all equivalent to 3SUM under subquadratic reductions. Finally, we show how to apply the methods of Fischer et al. to show that we can reduce non-trivial variant of 3LDT (Conv3LDT) over an arbitrary universe to the same variant over cubic (quadratic) universe.