The OEIS mourns the passing of Jim Simons and is grateful to the Simons Foundation for its support of research in many branches of science, including the OEIS.
login
The OEIS is supported by the many generous donors to the OEIS Foundation.

 

Logo
Hints
(Greetings from The On-Line Encyclopedia of Integer Sequences!)
A111000 Integer part of zeta(zeta(n)). 1

%I #27 Oct 24 2020 17:33:45

%S 2,5,12,27,58,120,245,498,1006,2024,4064,8149,16327,32692,65435,

%T 130938,261966,524051,1048260,2096731,4193743,8387860,16776219,

%U 33553102,67107091,134215365,268432305,536866711,1073736223,2147476181

%N Integer part of zeta(zeta(n)).

%C Notice the previous term is almost 1/2 the next term. Conjecture: lim_{n -> infinity} zeta(zeta(n))/zeta(zeta(n+1)) = 1/2.

%C If this sequence were defined instead as a(n) = lim_{k->n+} floor(zeta(zeta(k))), then it would be defined at n=1, with a(1) = 1. - _Iain Fox_, Sep 16 2020

%H Iain Fox, <a href="/A111000/b111000.txt">Table of n, a(n) for n = 2..3321</a>

%F Zeta(s) = Sum_{n >= 1} 1/n^s.

%F For n>=2, a(n) = floor(2^n-(4/3)^n-1+gamma+(8/9)^n-(4/5)^n+(2/3)^n). - _Benoit Cloitre_, Oct 04 2005

%F a(n) = 2^n-(4/3)^n+O(1) and more precisely lim_{n-->infinity} zeta(zeta(n))-2^n+(4/3)^n+1 = gamma where gamma is the Euler-Mascheroni constant. - _Benoit Cloitre_, Oct 04 2005

%F It appears that a(n) is always equal to or 1 greater than round(1/(zeta(n)-1)). - _Iain Fox_, Oct 27 2017 (edited Nov 22 2017)

%F From _Iain Fox_, Nov 22 2017: (Start)

%F For reference, b(n) = round(1/(zeta(n)-1)) and c(n) = ceiling(gamma(zeta(n)-1)).

%F a(n) >= b(n) >= c(n). By The Squeeze Theorem, if a(n) = c(n) for some integer n, then a(n) = b(n). (Proved by _Iain Fox_, Sep 16 2020)

%F a(n) - c(n) <= 1. (Proved by _Iain Fox_, Sep 16 2020)

%F (End)

%F Proof of above comments from _Iain Fox_, Sep 16 2020: (Start)

%F Consider a function f(x) = zeta(x) - 1/(x-1). Lim_{x->infinity} f(x) = 1 and lim_{x->1} f(x) = 0.577... is the Euler-Mascheroni constant. f(x) is strictly increasing, so the range of f(x) on the interval (1, infinity) is (0.577..., 1). a(n) is necessarily greater than or equal to b(n) when f(zeta(n)) >= 1/2. This is the case for n>1. Therefore, a(n) >= b(n).

%F Consider a function g(x) = 1/(x-1) - gamma(x-1). g(x) is greater than 0 strictly decreasing on the interval (1, 2). Lim_{x->1} g(x) = .577... is the Euler-Mascheroni constant. b(n) is necessarily greater than or equal to c(n) when g(zeta(n)) > 1/2. This is the case for n>3. Since b(2) = c(2) and b(3) = c(3), this proves b(n) >= c(n).

%F Consider a function h(x) = zeta(x) - gamma(x-1). Lim_{x->1} h(x) = 1.154... is twice the Euler-Mascheroni constant. On the interval (1, 2], h(x) is strictly decreasing and has values on the interval [Pi^2/6-1 = 0.644..., 1.154...). a(n) - c(n) is necessarily less than or equal to 1 when h(zeta(n)) < 2. This is the case for n>1. Therefore a(n) - c(n) <= 1.

%F (End)

%e a(100) ~ 1267650600228229398378720795167.

%e a(101) ~ 2535301200456458798836096530474.

%e a(100)/a(101) ~ 0.49999999999999999959005759..

%t IntegerPart[Zeta[Zeta[Range[2,40]]]] (* _Harvey P. Dale_, Sep 17 2019 *)

%o (PARI) zz(n) = for(x=2,n,print1(floor(zeta(zeta(x)))","))

%K easy,nonn

%O 2,1

%A _Cino Hilliard_, Sep 30 2005

Lookup | Welcome | Wiki | Register | Music | Plot 2 | Demos | Index | Browse | More | WebCam
Contribute new seq. or comment | Format | Style Sheet | Transforms | Superseeker | Recents
The OEIS Community | Maintained by The OEIS Foundation Inc.

License Agreements, Terms of Use, Privacy Policy. .

Last modified May 31 22:12 EDT 2024. Contains 373007 sequences. (Running on oeis4.)