OFFSET
1,1
LINKS
Charles R Greathouse IV, Table of n, a(n) for n = 1..10000
Mathematics Stack Exchange, crow, Can we prove B(n)=(1/4)G(n-1)G(n) is an indicator function ...
E. C. Titchmarsh, On van der Corput's Method and the zeta-function of Riemann IV, Quarterly Journal of Mathematics os-5 (1934), pp. 98-105.
Timothy Trudgian, On the success and failure of Gram's Law and the Rosser Rule, Acta Arithmetica, 2011 | 148 | 3 | 225-256.
Eric Weisstein's World of Mathematics, Gram Point.
FORMULA
Trudgian shows that a(n) = O(n), that is, there exists some k such that a(n) <= k*n. - Charles R Greathouse IV, Aug 29 2012
In fact Trudgian shows that a(n) ≍ n, and further, there exist constants 1 < b < c such that b*n < a(n) < c*n. (See the paper's discussion of the Weak Gram Law.) - Charles R Greathouse IV, Mar 28 2023
EXAMPLE
(-1)^126 Z(g(126)) = -0.0276294988571999.... - David Baugh, Apr 02 2008
MATHEMATICA
g[n_] := (g0 /. FindRoot[ RiemannSiegelTheta[g0] == Pi*n, {g0, 2*Pi*Exp[1 + ProductLog[(8*n + 1)/(8*E)]]}, WorkingPrecision -> 16]); Reap[For[n = 1, n < 1100, n++, If[(-1)^n*RiemannSiegelZ[g[n]] < 0, Print[n]; Sow[n]]]][[2, 1]] (* Jean-François Alcover, Oct 17 2012, after Eric W. Weisstein *)
PROG
(PARI) g0(n)=2*Pi*exp(1+lambertw((8*n+1)/exp(1)/8)) \\ approximate location of gram(n)
th(t)=arg(gamma(1/4+I*t/2))-log(Pi)*t/2 \\ theta, but off by some integer multiple of 2*Pi
thapprox(t)=log(t/2/Pi)*t/2-t/2-Pi/8+1/48/t-1/5760/t^3
RStheta(t)=my(T=th(t)); (thapprox(t)-T)\/(2*Pi)*2*Pi+T
gram(n)=my(G=g0(n), k=n*Pi); solve(x=G-.003, G+1e-8, RStheta(x)-k)
Z(t)=exp(th(t)*I)*zeta(1/2+I*t)
is(n)=my(G=gram(n)); real((-1)^n*Z(G))<0 \\ Charles R Greathouse IV, Jan 22 2022
CROSSREFS
KEYWORD
nonn
AUTHOR
Eric W. Weisstein, Jan 02 2006
EXTENSIONS
Definition corrected by David Baugh, Apr 02 2008
STATUS
approved