Settling the Optimal Exponent Relating Sumsets and Difference Sets
For finite sets A of integers, the sum-difference inequalities state σ(A)^{1/2} ≤ δ(A) ≤ σ(A)^2. The exponent 2 in the second inequality was known to be optimal. It was open whether the exponent 1/2 in the first inequality could be improved, equivalently whether sup log σ(A)/log δ(A) equals 2.
An explicit family A_K ⊂ ℤ is built from four ingredients: a base-12 digit gadget W={0,1,2,4,5,9}, a three-state carry automaton that counts modular differences, a symmetric additive basis in a cyclic group, and a Chinese remainder construction. The construction was found with the AI research agent Hyra (Hy3 model), guided by an LLM judge. The authors checked the proof by hand, and GPT-5.6 Sol converted the natural-language proof into Lean 4.
For every positive even K, C(A_K) > 2K/(K+3), so C(A_K) → 2 and the exponent 1/2 is optimal. This exceeds prior constructions and agent-search results, including Penman–Wells at 1.1259 and SimpleTES at 1.1449. A Lean 4 formal proof is released.
| Method | Value |
|---|---|
| Freiman–Pigarev (1973) | 1.0598 |
| Penman–Wells family | 1.1259 |
| AlphaEvolve | 1.1219 |
| SimpleTES with post-training | 1.1449 |
| Explicit family A_K (this work) | sup = 2 |
