English

Progress towards a nonintegrality conjecture

Number Theory 2019-03-20 v1

Abstract

Given rNr \in \mathbb{N}, define the function Sr:NQS_{r}: \mathbb{N} \rightarrow \mathbb{Q} by Sr(n)=k=0nkk+r(nk)S_{r}(n)=\displaystyle \sum_{k=0}^{n} \frac{k}{k+r} \binom{n}{k}. In 20152015, the second author conjectured that there are infinitely many rNr \in \mathbb{N} such that Sr(n)S_{r}(n) is nonintegral for all n1n \geq 1, and proved that Sr(n)S_{r}(n) is not an integer for r{2,3,4}r \in \{2,3,4\} and for all n1n \geq 1. In 20162016, Florian Luca and the second author raised the stronger conjecture that for any r1r \geq 1, Sr(n)S_{r}(n) is nonintegral for all n1n \geq 1. They proved that Sr(n)S_{r}(n) is nonintegral for r{5,6}r \in \{5,6\} and that Sr(n)S_{r}(n) is not an integer for any r2r \geq 2 and 1nr11 \leq n \leq r-1. In particular, for all r2r \geq 2, Sr(n)S_{r}(n) is nonintegral for at least r1r-1 values of nn. In 20182018, the fourth author gave sufficient conditions for the nonintegrality of Sr(n)S_{r}(n) for all n1n \geq 1, and derived an algorithm to sometimes determine such nonintegrality; along the way he proved that Sr(n)S_{r}(n) is nonintegral for r{7,8,9,10}r \in \{7,8,9,10\} and for all n1n \geq 1. By improving this algorithm we prove the conjecture for r22r\le 22. Our principal result is that Sr(n)S_r(n) is usually nonintegral in that the upper asymptotic density of the set of integers nn with Sr(n)S_r(n) integral decays faster than any fixed power of r1r^{-1} as rr grows.

Keywords

Cite

@article{arxiv.1903.08043,
  title  = {Progress towards a nonintegrality conjecture},
  author = {Shanta Laishram and Daniel López-Aguayo and Carl Pomerance and Thotsaphon Thongjunthug},
  journal= {arXiv preprint arXiv:1903.08043},
  year   = {2019}
}
R2 v1 2026-06-23T08:12:54.658Z