2
$\begingroup$

Let $s= 1/3 + 14i$. Prove that the real part of this limit converges to $\frac{1}{2}$:
$$ \Re\lim_{n \rightarrow \infty} \left( \left[ 1- \left( \sum _{k=1}^n \frac{(-1)^{k-1} \binom{n-1}{k-1}}{\zeta(\tfrac{k}{n}+s)} \Bigg/ \sum _{k=1}^n \frac{(-1)^{k-1} \binom{n-1}{k-1}}{\zeta(\tfrac{k}{n}+s+\tfrac{1}{n})} \right) \right]^{-1} +\frac1n + s \right) = \frac{1}{2}. $$ Mathematica 8.0.1:

n = 100;
s = (1/3 + 14*I);
s + 1/n + 
  1/(1 - Sum[(-1)^(k - 1)*Binomial[n - 1, k - 1]/Zeta[s + k/n], {k, 1,
         n}]/Sum[(-1)^(k - 1)*
        Binomial[n - 1, k - 1]/Zeta[s + k/n + 1/n], {k, 1, n}]);
N[%, n]

Output:
0.50000000000000000000000000000000000000000000000000000000000000000000
000000000000000000000000000000 +
14.134725141734693790457251983562470270784257115699243175685567460149\ 96342980925676494901039317156101 I

Pari GP:

n = 30
s = (1/3 + 14*I)
s + 1/n + 1/(1 - sum (k = 1, n, (-1)^(k - 1)*binomial(n - 1, k - 1)/zeta (s + k/n))/sum (k = 1, n, (-1)^(k - 1)*binomial(n - 1, k - 1)/zeta (s + k/n + 1/n)))

Copy Paste via mouse button and press Shift Enter in Pari GP to compute.
The working precision in Pari GP is not as good as in Mathematica. Therefore $n=30$ instead of $n=100$.

$\endgroup$
8
  • 8
    $\begingroup$ Perhaps the biggest typeset equation I have every seen. I wonder what the biggest equation ever published was. $\endgroup$
    – Ben McKay
    Commented Nov 7, 2020 at 16:34
  • 3
    $\begingroup$ Why do you need $1/n$ (and also why $+s$ in LHS but not $1/6$ on the right)? $\endgroup$ Commented Nov 7, 2020 at 17:40
  • 2
    $\begingroup$ It seems that if you start with an arbitrary $s$, your expression converges to a nearby zero of $\zeta$. Very curious (hence interesting). The first 100 billion nontrivial zeros of $\zeta$ are known to have real part $1/2$. See also plouffe.fr/simon/constants/zeta100.html $\endgroup$
    – GH from MO
    Commented Nov 7, 2020 at 18:26
  • 4
    $\begingroup$ This has nothing to do with zeta: try it with other functions, it gives very good approximations to a zero. $\endgroup$ Commented Nov 7, 2020 at 21:42
  • 1
    $\begingroup$ @GH: I have looked in standard refs, and did not find this formula. It is not very efficient, but it would be interesting to have some sort of proof. $\endgroup$ Commented Nov 9, 2020 at 12:26

4 Answers 4

1
$\begingroup$

Let: $$f(x)=\zeta (x)$$ $$A(n,s)=\sum _{k=1}^n \frac{(-1)^{k-1} \binom{n-1}{k-1}}{f\left(\frac{k}{n}-\frac{1}{n}+s\right)}$$ $$B(n,s)=\sum _{k=1}^n \frac{(-1)^{k-1} \binom{n-1}{k-1}}{f\left(\frac{k}{n}+s\right)}$$ $$X(n,s)=\sum _{k=1}^n \frac{(-1)^{k-1} \binom{n-1}{k-1}}{f\left(\frac{k}{n}-\frac{1}{n}-s\right)}$$ $$Y(n,s)=\sum _{k=1}^n \frac{(-1)^{k-1} \binom{n-1}{k-1}}{f\left(\frac{k}{n}-s\right)}$$ $$a=\frac{1}{1-\frac{A(n,s)}{B(n,s)}}+s$$ $$b=\frac{1}{1-\frac{B(n,s)}{A(n,s)}}-s$$ $$c=\frac{1}{1-\frac{X(n,s)}{Y(n,s)}}-s$$ Notice that: $$X(n,s)=A(n,-s)$$
and
$$Y(n,s)=B(n,-s)$$

For $s=1/3+14i$, show: $$\lim_{n\to \infty } \, ((a+b)(1-(b-c)))=1$$ Leaving out the limit symbol and substituting $a,b,c$: $$\left(\frac{1}{1-\frac{A}{B}}+\frac{1}{1-\frac{B}{A}}\right) \left(-\frac{1}{1-\frac{B}{A}}+\frac{1}{1-\frac{X}{Y}}+1\right)=1$$ which is equal to: $$-\frac{A Y+B X-2 B Y}{(A-B) (X-Y)}=1$$ multiplying with the denominator: $$-A Y-B X+2 B Y=(A-B) (X-Y)$$ subtracting with the right hand side: $$-A Y-B X+2 B Y -(A-B) (X-Y)=0$$ factoring: $$B Y-A X=0$$ which is: $$A X=B Y$$ Including the limit symbol again and substituting $A,B,X,Y$: $$\lim_{n\to \infty } \, \left(\left(\sum _{k=1}^n \frac{(-1)^{k-1} \binom{n-1}{k-1}}{f\left(\frac{k}{n}-\frac{1}{n}+s\right)}\right) \sum _{k=1}^n \frac{(-1)^{k-1} \binom{n-1}{k-1}}{f\left(\frac{k}{n}-\frac{1}{n}-s\right)}=\left(\sum _{k=1}^n \frac{(-1)^{k-1} \binom{n-1}{k-1}}{f\left(\frac{k}{n}+s\right)}\right) \sum _{k=1}^n \frac{(-1)^{k-1} \binom{n-1}{k-1}}{f\left(\frac{k}{n}-s\right)}\right)$$ For $n=30$ and $s=1/3+14i$ we get numerically:
for the left hand side:
$$55226.0411027488837442269063281-14296.8517199926101555805382701 i$$ and for the right hand side:
$$55226.0411027488837442269063281-14296.8517199926101555805382701 i$$ which appear close to each other.


while for example if we set the function to $f(x)=\zeta \left(x,\frac{1}{3}\right)$ we get numerically,
for the left hand side: $$-\text{5.7804095358568700751853287633386056719879460800172587438796357645$\grave{ }$30.044170983082427*${}^{\wedge}$-32}-\text{4.5948958062512910951997009159524637155472195553235558338875204958$\grave{ }$29.944488042052914*${}^{\wedge}$-32} i$$ and for the right hand side:
$$-\text{8.1747358863640979486289267057810389747002120618563412538892819063$\grave{ }$30.082847849900105*${}^{\wedge}$-32}-\text{4.9430723869496501619456415921898782802016090061312604239413643251$\grave{ }$29.864371090122273*${}^{\wedge}$-32} i$$ which are different.

Mathematica:

Clear[f, A, B, n, s, a, b, x, m];
f[x_] := Zeta[x];
A[n_, s_] := 
 Sum[(-1)^(k - 1)*Binomial[n - 1, k - 1]/f[s + k/n - 1/n], {k, 1, n}]
B[n_, s_] := 
 Sum[(-1)^(k - 1)*Binomial[n - 1, k - 1]/f[s + k/n], {k, 1, n}]
X[n_, s_] := 
 Sum[(-1)^(k - 1)*Binomial[n - 1, k - 1]/f[-s + k/n - 1/n], {k, 1, n}]
Y[n_, s_] := 
 Sum[(-1)^(k - 1)*Binomial[n - 1, k - 1]/f[-s + k/n], {k, 1, n}]

n = 30;
s = 1/3 + 14*I;
N[A[n, s]*X[n, s], 30]
N[B[n, s]*Y[n, s], 30]

Mathematica again:

Clear[f, A, B, n, s, a, b, x, m];
f[x_] = Zeta[x];
A[n_, s_] = 
  Sum[(-1)^(k - 1)*Binomial[n - 1, k - 1]/f[s + k/n - 1/n], {k, 1, n}];
B[n_, s_] = 
  Sum[(-1)^(k - 1)*Binomial[n - 1, k - 1]/f[s + k/n], {k, 1, n}];
X[n_, s_] = 
  Sum[(-1)^(k - 1)*Binomial[n - 1, k - 1]/f[-s + k/n - 1/n], {k, 1, 
    n}];
Y[n_, s_] = 
  Sum[(-1)^(k - 1)*Binomial[n - 1, k - 1]/f[-s + k/n], {k, 1, n}];
n = 120;
s = N[1 + 2*I , 200];
Block[{$MaxExtraPrecision = 600}, N[A[n, s]*X[n, s]/(B[n, s]*Y[n, s]), 20]]
N[1 + 5/((s/I)^2 + 4), 20]

which numerically suggests that:

$$\lim_{n\to \infty } \, \left(\frac{A(n,s) X(n,s)}{B(n,s) Y(n,s)} \right) =1+\frac{5}{4+\left(\frac{s}{i}\right)^2} $$ for those $s$ closest to the first trivial zero.

$\endgroup$
1
$\begingroup$

The limit in the question above can be found by simplifying ratios of consecutive repeated derivatives of $\frac{1}{\zeta(s)}$.


My recent thoughts:

See: https://mathoverflow.net/a/439430/25104

Let: $$\rho=\frac{1}{1-\frac{A}{B}}+\frac{1}{n}+s$$

Then according to Mathematica 8.0.1 the following:

$$\tag{1}$$ $$\text{Reduce}\left[\rho=\frac{1}{1-\frac{A}{B}}+\frac{1}{n}+s\land \frac{1}{1-\frac{A}{B}}+\frac{1}{n}+s=\left(-s+\frac{1}{1-\frac{B}{A}}-\frac{1}{n}\right)^*\land B\neq 0\land \Re(\rho)\geq 0\land \Re(\rho)\leq 1,\Re(\rho)\right]$$

reduces to the conditions:

$$\tag{2}$$ $A=\frac{B \cdot n \cdot \rho -B \cdot n \cdot s+B (-(n+1))}{n \cdot \rho-n \cdot s-1}$ $\land \left(\left(\Im(\rho)<\Im(s)\land \Re(\rho)=\frac{1}{2}\land (\Re(B)<0\lor (\Re(B)=0\land (\Im(B)<0\lor \Im(B)>0))\lor \Re(B)>0)\right)\lor \left(\Im(\rho)=\Im(s)\land \left(\left(\Re(s)<\frac{n-2}{2 n}\land \Re(\rho)=\frac{1}{2}\land (\Re(B)<0\lor (\Re(B)=0\land (\Im(B)<0\lor \Im(B)>0))\lor \Re(B)>0)\right)\lor \left(\Re(s)>\frac{n-2}{2 n}\land \Re(\rho)=\frac{1}{2}\land (\Re(B)<0\lor (\Re(B)=0\land (\Im(B)<0\lor \Im(B)>0))\lor \Re(B)>0)\right)\right)\right)\lor \left(\Im(\rho)>\Im(s)\land \Re(\rho)=\frac{1}{2}\land (\Re(B)<0\lor (\Re(B)=0\land (\Im(B)<0\lor \Im(B)>0))\lor \Re(B)>0)\right)\right)\land 0\leq \Re(\rho)\leq 1$

So in other words there are choices of $A$, $B$ and the starting point $s$ where $\Re(\rho) = 1/2$, but the conditions above must be satisfied in order for the following equation:

$$\lim_{n \rightarrow \infty}\left( \frac{1}{1-\frac{A}{B}}+\frac{1}{n}+s=\left(-s+\frac{1}{1-\frac{B}{A}}-\frac{1}{n}\right)^*\right)$$

to be true.

In the case of the Riemann zeta function we would write:

$$A=\sum _{k=1}^n \frac{(-1)^{k-1} \binom{n-1}{k-1}}{\zeta \left(\frac{k}{n}+s\right)} \tag{3}$$

$$B=\sum _{k=1}^n \frac{(-1)^{k-1} \binom{n-1}{k-1}}{\zeta \left(\frac{k}{n}+s+\frac{1}{n}\right)} \tag{4}$$

The parentheses with the star:
$$\left(\right)^*$$ is the complex conjugate.

How to show this in the case of the Riemann zeta function I don't know.

Demonstration in Mathematica 8.0.1:

(* Mathematica start *)
Clear[A, B, n, k, s];
n = 20;
s = 0;
A = Sum[(-1)^(k - 1)*Binomial[n - 1, k - 1]/Zeta[s + k/n], {k, 1, n}];
B = Sum[(-1)^(k - 1)*Binomial[n - 1, k - 1]/Zeta[s + k/n + 1/n], {k, 
    1, n}];
"The first trivial zero with the zero of the conjugated limit shifted \
by 1 and with opposite sign:"
Conjugate[-s - 1/n + 1/(1 - B/A)];
N[%, n]
 s + 1/n + 1/(1 - A/B);
N[%, n]

Clear[A, B, n, k, s];
n = 20;
s = (1/3 + 14*I);
A = Sum[(-1)^(k - 1)*Binomial[n - 1, k - 1]/Zeta[s + k/n], {k, 1, n}];
B = Sum[(-1)^(k - 1)*Binomial[n - 1, k - 1]/Zeta[s + k/n + 1/n], {k, 
    1, n}];
"The first non-trivial zero:"
Conjugate[-s - 1/n + 1/(1 - B/A)];
N[%, n]
 s + 1/n + 1/(1 - A/B);
N[%, n]

"The conditions to be proven for the case of the Riemann zeta function:"
Clear[s, A, B, n, z, k];
n = 19;
Reduce[rho == s + 1/n + 1/(1 - A/B) && 
  s + 1/n + 1/(1 - A/B) == Conjugate[-s - 1/n + 1/(1 - B/A)] && 
  B != 0 && Re[rho] >= 0 && Re[rho] <= 1, Re[rho]]
(*end*)
$\endgroup$
6
  • $\begingroup$ $s=14 i$ $$\lim_{n \rightarrow \infty} \left( \left[ 1- \left( \int _{-n}^{n} \frac{(-1)^{k-1} \binom{n-1}{k-1}}{\zeta(\tfrac{k}{n}+s)} \Bigg/ \int _{-n}^{n} \frac{(-1)^{k-1} \binom{n-1}{k-1}}{\zeta(\tfrac{k}{n}+s+\tfrac{1}{n})} \right) \right]^{-1} +\frac1n + s \right)=$$ $\endgroup$ Commented Feb 1, 2023 at 19:56
  • $\begingroup$ $$=\lim_{n \rightarrow \infty} \left( \left[ 1- \left( \sum _{-n}^n \frac{(-1)^{k-1} \binom{n-1}{k-1}}{\zeta(\tfrac{k}{n}+s)} \Bigg/ \sum _{-n}^n \frac{(-1)^{k-1} \binom{n-1}{k-1}}{\zeta(\tfrac{k}{n}+s+\tfrac{1}{n})} \right) \right]^{-1} +\frac1n + s \right)$$ $\endgroup$ Commented Feb 1, 2023 at 19:56
  • $\begingroup$ See Mathematica code at: chat.stackexchange.com/transcript/message/62894215#62894215 $\endgroup$ Commented Feb 1, 2023 at 20:08
  • $\begingroup$ "Mathematica start" Clear[s, A, B, n, z, k]; n = 19; Reduce[rho == s + 1/n + 1/(1 - A/B) && s + 1/n + 1/(1 - A/B) == Conjugate[-s - 1/n + 1/(1 - B/A)] && B != 0, Re[rho]] "Mathematica end" $\endgroup$ Commented Jul 26, 2023 at 16:16
  • $\begingroup$ "Mathematica start" Clear[s, A, B, n, z, k]; n = 100; s = (1/3 + 14*I); A = Sum[(-1)^(k - 1)*Binomial[n - 1, k - 1]/Zeta[s + k/n], {k, 1, n}]; B = Sum[(-1)^(k - 1)*Binomial[n - 1, k - 1]/Zeta[s + k/n + 1/n], {k, 1, n}]; {N[s + 1/n + 1/(1 - A/B), n], N[Conjugate[-s - 1/n + 1/(1 - B/A)], n]} "Mathematica end" $\endgroup$ Commented Jul 30, 2023 at 18:21
0
$\begingroup$

The starting point for deriving the formula in the question is the following limits for the derivatives of the Riemann zeta function:

$$\zeta '(s)=\lim_{c\to 1} \, \zeta (s) \left(\zeta (c)-\frac{\zeta (c) \zeta (s)}{\zeta (c+s-1)}\right)$$

$$\zeta ''(s)=\lim_{c\to 1} \, \zeta '(s) \left(\zeta (c)-\frac{\zeta (c) \zeta '(s)}{\zeta '(c+s-1)}\right)$$

$$\zeta ^{(3)}(s)=\lim_{c\to 1} \, \zeta ''(s) \left(\zeta (c)-\frac{\zeta (c) \zeta ''(s)}{\zeta ''(c+s-1)}\right)$$

$$\zeta ^{(4)}(s)=\lim_{c\to 1} \, \zeta ^{(3)}(s) \left(\zeta (c)-\frac{\zeta (c) \zeta ^{(3)}(s)}{\zeta ^{(3)}(c+s-1)}\right)$$

$$\zeta ^{(5)}(s)=\lim_{c\to 1} \, \zeta ^{(4)}(s) \left(\zeta (c)-\frac{\zeta (c) \zeta ^{(4)}(s)}{\zeta ^{(4)}(c+s-1)}\right)$$

$$\zeta ^{(6)}(s)=\lim_{c\to 1} \, \zeta ^{(5)}(s) \left(\zeta (c)-\frac{\zeta (c) \zeta ^{(5)}(s)}{\zeta ^{(5)}(c+s-1)}\right)$$

$$\vdots$$

$$\zeta ^{(n+1)}(s)=\lim_{c\to 1} \, \zeta ^{(n)}(s) \left(\zeta (c)-\frac{\zeta (c) \zeta ^{(n)}(s)}{\zeta ^{(n)}(c+s-1)}\right)$$

In mathematica this would be:

Clear[s, c]
Limit[((-Zeta[s]*Zeta[c]/Zeta[s + c - 1] + Zeta[c])*Zeta[s]), c -> 1]
Limit[((-Zeta'[s]*Zeta[c]/Zeta'[s + c - 1] + Zeta[c])*Zeta'[s]), c -> 1]
Limit[((-Zeta''[s]*Zeta[c]/Zeta''[s + c - 1] + Zeta[c])*Zeta''[s]), c -> 1]
Limit[((-Zeta'''[s]*Zeta[c]/Zeta'''[s + c - 1] + Zeta[c])*Zeta'''[s]), c -> 1]
Limit[((-Zeta''''[s]*Zeta[c]/Zeta''''[s + c - 1] + Zeta[c])*Zeta''''[s]), c -> 1]
Limit[((-Zeta'''''[s]*Zeta[c]/Zeta'''''[s + c - 1] + Zeta[c])*Zeta'''''[s]), c -> 1]

It is known that the ratios of consecutive derivatives a function in general converges to the nearest root (or singularity).

So: $$\lim\limits_{n \rightarrow \infty} \frac{\zeta ^{(n+1)}(s)}{\zeta ^{(n)}(s)} + s = \text{nearest zero of the Riemann zeta function}$$

By succesively substituting the right hand side of the lower order derivatives of Riemann zeta into the right hand side of the immediate higher order derivative of Riemann zeta, with the help of the following Mathematica program;

Clear[s, c, A]
A0 = 1/Zeta[s];
Limit[Zeta[c] A0 - Zeta[c]/Zeta[-1 + c + s], c -> 1];

A1 = Zeta[c]/Zeta[-0 + 0 c + s] - Zeta[c]/Zeta[-1 + 1 c + s];
A2 = Zeta[c]/Zeta[-1 + 1 c + s] - Zeta[c]/Zeta[-2 + 2 c + s];
A3 = Zeta[c]/Zeta[-2 + 2 c + s] - Zeta[c]/Zeta[-3 + 3 c + s];
A4 = Zeta[c]/Zeta[-3 + 3 c + s] - Zeta[c]/Zeta[-4 + 4 c + s];
A5 = Zeta[c]/Zeta[-4 + 4 c + s] - Zeta[c]/Zeta[-5 + 5 c + s];

B1 = ReplaceAll[A1, Zeta[-1 + 1 c + s] -> 1/A2];
B2 = ReplaceAll[B1, Zeta[-0 + 0 c + s] -> 1/A1];

C1 = ReplaceAll[B2, Zeta[-2 + 2 c + s] -> 1/A3];
C2 = ReplaceAll[C1, Zeta[-1 + 1 c + s] -> 1/A2];
C3 = ReplaceAll[C2, Zeta[-0 + 0 c + s] -> 1/A1];

D1 = ReplaceAll[C3, Zeta[-3 + 3 c + s] -> 1/A4];
D2 = ReplaceAll[D1, Zeta[-2 + 2 c + s] -> 1/A3];
D3 = ReplaceAll[D2, Zeta[-1 + 1 c + s] -> 1/A2];
D4 = ReplaceAll[D3, Zeta[-0 + 0 c + s] -> 1/A1];

E1 = ReplaceAll[D4, Zeta[-4 + 4 c + s] -> 1/A5];
E2 = ReplaceAll[E1, Zeta[-3 + 3 c + s] -> 1/A4];
E3 = ReplaceAll[E2, Zeta[-2 + 2 c + s] -> 1/A3];
E4 = ReplaceAll[E3, Zeta[-1 + 1 c + s] -> 1/A2];
E5 = ReplaceAll[E4, Zeta[-0 + 0 c + s] -> 1/A1];

FullSimplify[A0]
FullSimplify[A1]
FullSimplify[B2]
FullSimplify[C3]
FullSimplify[D4]
FullSimplify[E5]

one gets:

FullSimplify[A0] $$\frac{1}{\zeta (s)}$$ FullSimplify[A1] $$\zeta (c) \left(\frac{1}{\zeta (s)}-\frac{1}{\zeta (c+s-1)}\right)$$ FullSimplify[A2] $$\zeta (c)^2 \left(\frac{1}{\zeta (s)}-\frac{2}{\zeta (c+s-1)}+\frac{1}{\zeta (2 c+s-2)}\right)$$ FullSimplify[A3] $$\zeta (c)^3 \left(\frac{1}{\zeta (s)}-\frac{3}{\zeta (c+s-1)}+\frac{3}{\zeta (2 c+s-2)}-\frac{1}{\zeta (3 c+s-3)}\right)$$ FullSimplify[A4] $$\zeta (c)^4 \left(\frac{1}{\zeta (s)}-\frac{4}{\zeta (c+s-1)}+\frac{6}{\zeta (2 c+s-2)}-\frac{4}{\zeta (3 c+s-3)}+\frac{1}{\zeta (4 c+s-4)}\right)$$ FullSimplify[A5] $$\zeta (c)^5 \left(\frac{1}{\zeta (s)}-\frac{5}{\zeta (c+s-1)}+\frac{10}{\zeta (2 c+s-2)}-\frac{10}{\zeta (3 c+s-3)}+\frac{5}{\zeta (4 c+s-4)}-\frac{1}{\zeta (5 c+s-5)}\right)$$

That looks like the binomial coefficients, and should be because of the binary(?) substition.

Imitating the last formula:

Clear[n, k, c, h, s]
n = 6;
Table[(-1)^(k + 1)*
  Binomial[n - 1, k - 1]/Zeta[-(k - 1) + (k - 1) c + s], {k, 1, n}]

So:

$$\frac{\zeta (c)^{n-1} \sum _{k=1}^n \frac{(-1)^{k+1} \binom{n-1}{k-1}}{\zeta (c (k-1)-(k-1)+s)}}{\zeta (c)^{n+1-1} \sum _{k=1}^{n+1} \frac{(-1)^{k+1} \binom{n+1-1}{k-1}}{\zeta (c (k-1)-(k-1)+s)}} \tag{1}$$

One can strike out part of the binomials to get the $\frac{A}{B}$ formula in the question.

How ever, in order to not lose information, one sets $c \rightarrow 1+\frac{1}{h}$

Because both $n \rightarrow \infty$ and $h \rightarrow \infty$ and we want to have a formula with as few variables as possible, we simply replace $h$ with $n$ (that is, $h=n$ from now on, because we say so). So:

$$c=1+\frac{1}{n}$$

close to $c=1$: $\zeta (c) \approx \frac{1}{c-1}$

introducing these changes, except striking out binomials, into $(1)$ we get:

$$\frac{\sum _{k=1}^n \frac{(-1)^{k+1} \binom{n-1}{k-1}}{\zeta \left(\frac{k-1}{n}+s\right)}}{\zeta \left(1+\frac{1}{n}\right) \left(\sum _{k=1}^{n+1} \frac{(-1)^{k+1} \binom{n}{k-1}}{\zeta \left(\frac{k-1}{n}+s\right)}\right)}$$

Striking out binomials we get:

$$\Re\lim_{n \rightarrow \infty} \left( \left[ 1- \left( \sum _{k=1}^n \frac{(-1)^{k-1} \binom{n-1}{k-1}}{\zeta(\tfrac{k}{n}+s)} \Bigg/ \sum _{k=1}^n \frac{(-1)^{k-1} \binom{n-1}{k-1}}{\zeta(\tfrac{k}{n}+s+\tfrac{1}{n})} \right) \right]^{-1} \underset{\text{conjectured guesses}}{\underbrace{+\frac1n + s}} \right) = \frac{1}{2}.$$

with conjectured terms $$+\frac1n + s$$ that are guesses to make it work.

We now have the formula in the form:

$$\rho=\frac{1}{1-\frac{A}{B}}+\frac{1}{n}+s$$

Please go to the previous comment as an answer: https://mathoverflow.net/a/377346/25104

$\endgroup$
1
0
$\begingroup$

It appears that we don't need analytic continuation to locate the first non-trivial Riemann zeta zero:

(*start*)
(*Mathematica 8.0.1*)
n = 100;(*set n=200 for more digits*)
s = (3/2 + 14*I);
s + 1/n + 
  1/(1 - Sum[(-1)^(k - 1)*
        Binomial[n - 1, k - 1]/HarmonicNumber[10^10000, s + k/n], {k, 
        1, n}]/Sum[(-1)^(k - 1)*
        Binomial[n - 1, k - 1]/
         HarmonicNumber[10^10000, s + k/n + 1/n], {k, 1, n}]);
N[%, n]
(*end*)

Output:

0.49999999999999999999999999999999999999999999999999999999999999999999\
821263876336076842104900392907 + 
 14.134725141734693790457251983562470270784257115699243175685567460149\
95997175784113908425467589217131 I
$\endgroup$
3
  • $\begingroup$ $s= 3/2 + 14i \\ M=10^{10000}$ $$\Re\lim_{n \rightarrow 100} \left( \left[ 1- \left( \sum _{k=1}^n \frac{(-1)^{k-1} \binom{n-1}{k-1}}{\sum_{m=1}^{m=M} 1/m^{\left(\tfrac{k}{n}+s\right)}} \Bigg/ \sum _{k=1}^n \frac{(-1)^{k-1} \binom{n-1}{k-1}}{\sum_{m=1}^{m=M} 1/m^{\left(\tfrac{k}{n}+s+\tfrac{1}{n}\right)}} \right) \right]^{-1} +\frac1n + s \right) = $$ 0.49999999999999999999999999999999999999999999999999999999999999999999\ 821263876336076842 + 14.134725141734693790457251983562470270784257115699243175685567460149\ 959971757841139084 I $\endgroup$ Commented Feb 13, 2023 at 14:42
  • $\begingroup$ This is analytic continuation after all, although I did not think so at first. $\endgroup$ Commented Mar 19, 2023 at 8:40
  • $\begingroup$ (*start*)(*Mathematica 8.0.1*) n = 40;(*set n=40 for more digits*) s = (1 + 14*I); x = 1/10^20; N[-s - 1/n - 1/(1 - Sum[(-1)^(k - 1)* Binomial[n - 1, k - 1]/((s + k/n) Zeta[1 + s + k/n] x), {k, 1, n}]/Sum[(-1)^(k - 1)* Binomial[n - 1, k - 1]/((s + k/n + 1/n) Zeta[1 + s + k/n + 1/n] x), {k, 1, n}]), n/2](*end*) $\endgroup$ Commented Aug 20, 2023 at 20:39

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .