The quadratic Symmetric Gaussian Problem

About 11 years old · traced to

Let nn be a positive integer, let 0<a<10<a<1, and let −1≤ρ≤1-1\leq\rho\leq1. For a measurable set A⊆RnA\subseteq\mathbb{R}^{n} with γn(A)=a\gamma_n(A)=a, choose ra,ra′>0r_a,r_a'>0 such that

γn(B(0,ra))=a,γn(B(0,ra′)c)=a.\gamma_n(B(0,r_a))=a,\qquad \gamma_n(B(0,r_a')^c)=a.

Quadratic Symmetric Gaussian Problem. Either B(0,ra)B(0,r_a) or B(0,ra′)cB(0,r_a')^c achieves

sup⁡A⊆Rn:γn(A)=a,A=−A∫Rn1A(x)Tρ(1A)(x) dγn(x).\sup_{\substack{A\subseteq\mathbb{R}^{n}:\\ \gamma_n(A)=a,\,A=-A}}\int_{\mathbb{R}^{n}}1_A(x)T_\rho(1_A)(x)\,d\gamma_n(x).

This is the special case obtained by restricting the Symmetric Gaussian Problem to a+b=1a+b=1 and A=BcA=B^c; the source provides no evidence that this quadratic version has been resolved.

References

Primary source

Steven Heilman, “Low Correlation Noise Stability of Symmetric Sets”, arXiv:1511.00382 (2016).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.