OctoInformBeta

The Gaussian propeller conjecture in every dimension

For positive integers dd and kk, let γd\gamma_d denote standard Gaussian measure on ℝd\mathbb R^d. The Gaussian first moment of a measurable set A⊆ℝdA\subseteq\mathbb R^d is the unnormalized vector z(A)=∫Axdγd(x)z(A)=\int_A x\,d\gamma_d(x). Let 𝒜=(A1,…,Ak)\mathcal A=(A_1,\ldots,A_k) be a measurable partition of ℝd\mathbb R^d up to Gaussian null sets, without prescribed cell probabilities. Empty cells are permitted. Its squared-moment value is ℱ(𝒜)=∑i=1k∥z(Ai)∥2\mathcal F(\mathcal A)=\sum_{i=1}^k\|z(A_i)\|^2. The propeller partition consists of three planar sectors of angle 2π/32\pi/3, extended by an orthogonal Euclidean factor, with any additional cells empty. The propeller conjecture asks whether this partition maximizes ℱ\mathcal F whenever d≥2d\ge2 and k≥3k\ge3.

The preprint proves ℱ(𝒜)≤9/(8π)\mathcal F(\mathcal A)\le9/(8\pi) in every positive dimension for every finite measurable partition. The propeller partition attains equality when d≥2d\ge2 and k≥3k\ge3. The preprint also derives a sharp inequality for the expected maximum of a finite centered Gaussian vector.

The Lean library includes the partition bound for cells measurable for the completion of Gaussian measure, assuming unique cell membership almost everywhere. Lean has checked every result of the Lean library, so the Lean library is the reference for the statements on these pages. Each result is stated as the Lean library states it and linked to the corresponding Lean declaration. Where a statement differs from the preprint, the Lean library’s statement is given with the preprint’s beside it. The preprint is a guide to the results’ order and names.