[Paper Review] A new expander and improved bounds for $A(A+A)$

Oliver Roche‐Newton|arXiv (Cornell University)|Mar 22, 2016

Limits and Structures in Graph Theory7 references3 citations

TL;DR

This paper introduces a novel five-variable expander function, $(a_1+a_2+a_3+a_4)^2 + \log a_5$, proving that its image size is at least $\gg |A|^2 / \log|A|$, which is optimal up to logarithmic factors. It further improves the best-known lower bounds for $|A(A-A)|$ and $|A(A+A)|$, achieving $\gtrapprox |A|^{3/2 + 1/34}$ and $\gtrapprox |A|^{3/2 + 5/242}$, respectively, using refined energy-based methods and avoiding the Balog-Szemerédi-Gowers theorem.

ABSTRACT

The main result in this paper concerns a new five-variable expander. It is proven that for any finite set of real numbers $A$, $$|\{(a_1+a_2+a_3+a_4)^2+\log a_5 :a_1,a_2,a_3,a_4,a_5 \in A \}| \gg \frac{|A|^2}{\log |A|}.$$ This bound is optimal, up to logarithmic factors. The paper also gives new lower bounds for $|A(A-A)|$ and $|A(A+A)|$, improving on results from arXiv:1312.6438. The new bounds are $$|A(A-A)| \gtrapprox |A|^{3/2+\frac{1}{34}}$$ and $$|A(A+A)| \gtrapprox |A|^{3/2+\frac{5}{242}}.$$

Motivation & Objective

• To establish a new optimal five-variable expander function involving sum of four elements squared and logarithm of a fifth.
• To improve the known lower bounds for the product sets $A(A-A)$ and $A(A+A)$ in the context of sum-product theory.
• To refine existing sum-product estimates by avoiding the Balog-Szemerédi-Gowers theorem and using energy decomposition techniques.
• To achieve quantitative improvements in the exponent of the growth rate for $|A(A-A)|$ and $|A(A+A)|$ beyond previous results.

Proposed method

• Introduces a new five-variable expander: $f(a_1,a_2,a_3,a_4,a_5) = (a_1+a_2+a_3+a_4)^2 + \log a_5$, and proves its image size is $\gg |A|^2 / \log|A|$.
• Applies dyadic pigeonholing to multiplicative energy $E^*(A)$, decomposing the set $A$ into subsets with controlled intersection sizes with scaled sets.
• Uses the structure of multiplicative energy and representation functions to bound the size of $A(A-A)$ and $A(A+A)$ via energy decomposition and dual energy estimates.
• Employs Lemma 4.2 and Lemma 4.3 to relate the size of difference and sum sets to the multiplicative energy, avoiding the use of Balog-Szemerédi-Gowers.
• Applies case analysis based on the multiplicative energy parameter $K = |A|^3 / E^*(A)$, splitting into high and low $K$ regimes to derive optimal bounds.
• Uses the inequality $|A(A-A)| \gtrapprox |A|^{3/2 + 1/34}$ and $|A(A+A)| \gtrapprox |A|^{3/2 + 5/242}$, derived from energy-based lower bounds on $|A-A|$ and $|A+A|$.

Experimental results

Research questions

• RQ1 Can a new five-variable expander function be constructed that achieves optimal growth rate up to logarithmic factors?
• RQ2 What is the best possible lower bound for $|A(A-A)|$ in terms of $|A|$ for finite sets of real numbers?
• RQ3 How can the sum-product bounds for $|A(A+A)|$ be improved beyond the state-of-the-art?
• RQ4 Can the Balog-Szemerédi-Gowers theorem be avoided in deriving such bounds while maintaining tightness?

Key findings

• The image of the five-variable function $(a_1+a_2+a_3+a_4)^2 + \log a_5$ has size $\gg |A|^2 / \log|A|$, which is optimal up to logarithmic factors.
• The bound $|A(A-A)| \gtrapprox |A|^{3/2 + 1/34}$ improves upon the previous $|A|^{3/2 + 1/112}$ from [8].
• The bound $|A(A+A)| \gtrapprox |A|^{3/2 + 5/242}$ improves upon the prior $|A|^{3/2 + 1/178}$ from [8].
• The method avoids the Balog-Szemerédi-Gowers theorem, streamlining the argument via energy decomposition and dyadic pigeonholing.
• The proof splits into two cases based on the multiplicative energy parameter $K = |A|^3 / E^*(A)$, achieving tight bounds in both regimes.
• The results are tight up to logarithmic factors, as confirmed by testing against arithmetic progressions.