Is it true that, for every finite p \geq 1, we have e(Tpn ) \leq Cp n?
Statement
Is it true that, for every finite p \geq 1, we have e(Tpn ) \leq Cp n? Can it be true that moreover, if 2n + 1 is a prime, then e(Tpn ) = 2n + 1? Hans Yu The following question is motivated by incidence geometry. What is the maximum number of vertex sets S of size r such that H[S] contains at least one edge of each color? The answer to this question is monotonically non-decreasing in r. \sqrt{ } Theorem 9. Noga Alon Definition 6. A set in a metric space X is equilateral if all distances between pairs are the same. Write e(X) = sup{|A| : A equilateral}. There are many conjectures about e(X) for normed spaces X, and very little is known. We do not even know e(ℓn1 ). Kravitz and Steinerberger suggest consideration of the following family of metric spaces. Write Tpn for the n-torus [0, 1)n equipped with the ℓp distance n !1/p X d(a, b) = min(|ai − bi |, 1 − |ai − bi |)p . i=1 One always has e(Tpn ) \geq 2n; this is given by the set {\pm\alphae1 , . . . , \pm\alphaen } where e1 , . . . , en are the elementary vectors and \alpha is chosen appropriately. If 2n + 1 is prime, then in fact e(Tpn ) \geq 2n + 1. For p = \infty we have e(T\infty n ) = 3n . The best upper bound known for fixed p < \infty is Cp n2+2/⌊p⌋ .
Record
Comments
Nothing has been published against this problem yet, and nothing has been said about it.
Sign in with an institutional address to take part in the discussion. Reading every thread stays open to everyone.
Sign inSolve with an agent
Open the statement in a chat, with the problem and the ground rules already written into the prompt.