2025/10/16 by Gandhi, Anshula
#Combinatorics (math.CO) #FOS: Mathematics #Metric Geometry (math.MG)
paper · doi:10.48550/arxiv.2510.14585
The distinct dot products problem, a variation on the Erdős distinct distance problem, asks "Given a set Pn of n points in ℝ2, what is the minimum number |D(Pn)| of distinct dot products formed between them, asymptotically?" The best proven lower-bound is |D(Pn)| \gtrsim n2/3+7/1425, due to work by Hanson\unicodex2013Roche-Newton\unicodex2013Senger, and a recent improvement by Kokkinos. However, the slowest-scaling known constructions have |D(Pn)|∼ n, leaving quite a large gap in the bound. Finding a sublinearly-scaling construction, or disproving its existence, would narrow this gap. We provide a condition that a sequence of point configurations (Pn)n ∈ ℕ must satisfy in order for |D(Pn)| to scale 'slowly' i.e. |D(Pn)| ≪ n3/4. Namely, we prove that any such configuration must contain a point-rich line that gets arbitrarily 'dense' as the sequence progresses.