English

Zerofree region for exponenetial sums

Number Theory 2011-06-17 v1

Abstract

We consider the following two closed sets in CnC^n. One is the diagonal D given by (z,z,z,...z) (z, z, z,...z_). The other is A={(z1,z2,z3,...zn):.A = \{(z_1,z_2,z_3,...z_n):. .ez1+ez2+ez3+...+ezn=0}.e^{z_1} + e^{z_2} +e^{z_3} +...+ e^{z_n}=0\}. Clearly DAD \cap A is empty. One can ask what is the distance between them. In this connection, Stolarsky [1] proved that the distance dd is given by d2=(log n)2+O(1)d^2 = (\log \ n)^2 + O (1). Some simple calculations will make one believe that the point (k,0,0,..0)(k, 0, 0,..0) with k=log (n1)+πik = \log \ (n-1) + \pi i which lies on AA is one of the closest point to the diagaonal. We prove that this is indeed the case, atleast for sufficiently large nn. This gives d2=k2(11/n)d^2 = |k|^2 (1-1/n).

Keywords

Cite

@article{arxiv.1106.3280,
  title  = {Zerofree region for exponenetial sums},
  author = {R. Balasubramanian},
  journal= {arXiv preprint arXiv:1106.3280},
  year   = {2011}
}
R2 v1 2026-06-21T18:23:27.916Z