Esta secuencia A177885 en la oeis aparentemente se relaciona imaginaria de la no-trivial de Riemann zeta ceros con el LambertW función. Las partes real e imaginaria de la de Riemann zeta función es la suma de las ondas seno y coseno con logaritmos como frecuencias.
Los logaritmos se puede calcular como:
$$\log(n)=\lim_{s\to 1} \, \left(1-\frac{1}{n^{s-1}}\right) \zeta (s)$$
of which the numerators in the Dirichlet series are found in the following infinite table:
$$T = \begin{bmatrix} 0&0&0&0&0&0&0 \\ 1&-1&1&-1&1&-1&1 \\ 1&1&-2&1&1&-2&1 \\ 1&1&1&-3&1&1&1 \\ 1&1&1&1&-4&1&1 \\ 1&1&1&1&1&-5&1 \\ 1&1&1&1&1&1&-6 \end{bmatrix}$$
which has the definition:
$$T(n,k) = -(n-1)\; \text{ if }\; n|k, \;\text{ else } \;1,$$
Repeating/recursing the formula above we write:
$$\log(a(n))= \lim_{s\to 1} \, \zeta (s) \sum _{k=1}^n \frac{T(n,k)}{k^{s-1}}$$
where a(n) appears to be: $$a(n)=\frac{n^n}{n!}$$
$a(n) =$ {1, 2, 9/2, 32/3, 625/24, 324/5, 117649/720, 131072/315, 4782969/4480, 1562500/567, 25937424601/3628800, 35831808/1925,...}
$\left\{1,2,\frac{9}{2},\frac{32}{3},\frac{625}{24},\frac{324}{5},\frac{117649}{720},\frac{131072}{315},\frac{4782969}{4480},\frac{1562500}{567},\frac{25937424601}{3628800},\frac{35831808}{1925}\right\}$
multiplicando con el factorial se encuentra el similar pero alternando la secuencia A177885 en la oeis.
Hay en el comentario de esta fórmula aproximada es:
Table[N[1/2 + 2*Pi*Exp[1]*(n - 11/8)/Exp[1]/LambertW[(n - 11/8)/Exp[1]]*I], {n,
1, 12}]
Table[N[ZetaZero[n]], {n, 1, 12}]
lo que da:
{0.5 + 14.5213 I, 0.5 + 20.6557 I, 0.5 + 25.4927 I, 0.5 + 29.7394 I,
0.5 + 33.6245 I, 0.5 + 37.2574 I, 0.5 + 40.7006 I, 0.5 + 43.994 I,
0.5 + 47.1651 I, 0.5 + 50.2337 I, 0.5 + 53.2144 I, 0.5 + 56.1189 I}
{0.5 + 14.1347 I, 0.5 + 21.022 I, 0.5 + 25.0109 I, 0.5 + 30.4249 I,
0.5 + 32.9351 I, 0.5 + 37.5862 I, 0.5 + 40.9187 I, 0.5 + 43.3271 I,
0.5 + 48.0052 I, 0.5 + 49.7738 I, 0.5 + 52.9703 I, 0.5 + 56.4462 I}
La Serie de x/LambertW es:
Series[x/LambertW[x], {x, 0, 7}]
$$\frac{x}{W(x)} = 1+x-\frac{x^2}{2}+\frac{2 x^3}{3}-\frac{9 x^4}{8}+\frac{32 x^5}{15}-\frac{625 x^6}{144}+\frac{324 x^7}{35}+O\left(x^8\right)$$
which has some similarity with $a(n)$
$$\frac{x}{W(x)} = \frac{(-1)^n n^n x^{n+1}}{(n+1)!}$$
$$a(n)=\frac{n^n}{n!}$$
$\left\{\frac{1}{2},\frac{2}{3},\frac{9}{8},\frac{32}{15},\frac{625}{144},\frac{324}{35},\frac{117649}{5760},\frac{131072}{2835},\frac{4782969}{44800},\frac{1562500}{6237},\frac{25937424601}{43545600},\frac{35831808}{25025}\right\}$
Is there a connection?
Edit 7.9.2013:
Would these sequences give more accurate power series approximations? Just a thought.
Clear[t, s, nn, m, k, n];
m = 1;
nn = 12;
t[n_, 1] = 1;
t[1, k_] = 1;
t[n_, k_] := t[n, k] = (1 - If[Mod[k, n] == 0, n, 0]);
MatrixForm[Table[Table[t[n, k], {k, 1, m*nn}], {n, 1, m*nn}]];
Print["here"]
Monitor[A =
Table[Limit[Zeta[s]*Sum[t[n, k]/k^(s - 1), {k, 1, m*n}],
s -> 1], {n, 1, nn}], n]
Clear[t, s, nn, m, k, n];
m = 2;
nn = 12;
t[n_, 1] = 1;
t[1, k_] = 1;
t[n_, k_] := t[n, k] = (1 - If[Mod[k, n] == 0, n, 0]);
MatrixForm[Table[Table[t[n, k], {k, 1, m*nn}], {n, 1, m*nn}]];
Print["here"]
Monitor[A =
Table[Limit[Zeta[s]*Sum[t[n, k]/k^(s - 1), {k, 1, m*n}],
s -> 1], {n, 1, nn}], n]
Clear[t, s, nn, m, k, n];
m = 3;
nn = 12;
t[n_, 1] = 1;
t[1, k_] = 1;
t[n_, k_] := t[n, k] = (1 - If[Mod[k, n] == 0, n, 0]);
MatrixForm[Table[Table[t[n, k], {k, 1, m*nn}], {n, 1, m*nn}]];
Print["here"]
Monitor[A =
Table[Limit[Zeta[s]*Sum[t[n, k]/k^(s - 1), {k, 1, m*n}],
s -> 1], {n, 1, nn}], n]
Clear[t, s, nn, m, k, n];
m = 4;
nn = 12;
t[n_, 1] = 1;
t[1, k_] = 1;
t[n_, k_] := t[n, k] = (1 - If[Mod[k, n] == 0, n, 0]);
MatrixForm[Table[Table[t[n, k], {k, 1, m*nn}], {n, 1, m*nn}]];
Print["here"]
Monitor[A =
Table[Limit[Zeta[s]*Sum[t[n, k]/k^(s - 1), {k, 1, m*n}],
s -> 1], {n, 1, nn}], n]
Clear[t, s, nn, m, k, n];
m = 5;
nn = 12;
t[n_, 1] = 1;
t[1, k_] = 1;
t[n_, k_] := t[n, k] = (1 - If[Mod[k, n] == 0, n, 0]);
MatrixForm[Table[Table[t[n, k], {k, 1, m*nn}], {n, 1, m*nn}]];
Print["here"]
Monitor[A =
Table[Limit[Zeta[s]*Sum[t[n, k]/k^(s - 1), {k, 1, m*n}],
s -> 1], {n, 1, nn}], n]
Edit 7.9.2013:
The connection I was looking for:
$$\sum _{n=1}^{\infty} \frac{x (-x)^n \exp \left(\lim_{s\to 1} \, \zeta (s) \sum _{k=1}^n \frac{1-\text{If}[k \bmod n=0,n,0]}{k^{s-1}}\right)}{n+1}+x+1 =1+x-\frac{x^2}{2}+\frac{2 x^3}{3}-\frac{9 x^4}{8}+\frac{32 x^5}{15}-\frac{625 x^6}{144}+\frac{324 x^7}{35}-\frac{117649 x^8}{5760}+\frac{131072 x^9}{2835}-\frac{4782969 x^{10}}{44800}+\frac{1562500 x^{11}}{6237}-\frac{25937424601 x^{12}}{43545600}+\frac{35831808 x^{13}}{25025}-...$$
1 + x + Sum[
x*(-x)^n*Exp[
Limit[Zeta[s]*
Sum[(1 - If[Mod[k, n] == 0, n, 0])/k^(s - 1), {k, 1, n}],
s -> 1]]/(n + 1), {n, 1, 12}]
Series[x/LambertW[x], {x, 0, 12}]
Edit 2.10.2013: Integration is better:
Clear[x, n, k, s, a1, nn, b1]
b1 = Expand[
Sum[Exp[Limit[
1/(s - 1)*
Sum[(1 - If[Mod[k, n] == 0, n, 0])/(k)^(s - 1), {k, 1, 4*n}],
s -> 1]]*(-x)^n, {n, 0, 32}]];
a1 = 1 + Integrate[b1, x];
x = N[(1 - 11/8)/Exp[1], 30];
Print["here"]
N[2*Pi*Exp[1]*a1, 30]
N[2*Pi*Exp[1]*x/LambertW[x], 30]
Clear[x, n, k, s, a1, nn]
a1 = 1 + Integrate[b1, x];
x = N[(2 - 11/8)/Exp[1], 30];
Print["here"]
N[2*Pi*Exp[1]*a1, 30]
N[2*Pi*Exp[1]*x/LambertW[x], 30]
where the number $4$ within: {k, 1, 4*n}], can be varied for truncating the Dirichlet series for the logarithm of $n$. Al menos tan largo como el truncado de la serie de Dirichlet no obtener más de la potencia de serie, hay una tendencia para que los Zeta cero aproximaciones a permanecer cerca de la zeta ceros.
12.10.2013: Una mejor integración:
Clear[x, n, k, s, a1, nn, b1]
b1 = Expand[
Sum[Exp[Limit[
Zeta[s]*Sum[(1 - If[Mod[k, n] == 0, n, 0])/k^(s - 1), {k, 1,
n}], s -> 1]]*(-x)^n, {n, 1, 32}]];
a1 = 1 + Integrate[1 + b1, x];
x = N[(1 - 11/8)/Exp[1], 30];
Print["here"]
N[2*Pi*Exp[1]*a1, 30]
N[2*Pi*Exp[1]*x/LambertW[x], 30]
Clear[x, n, k, s, a1, nn]
a1 = 1 + Integrate[1 + b1, x];
x = N[(2 - 11/8)/Exp[1], 30];
Print["here"]
N[2*Pi*Exp[1]*a1, 30]
N[2*Pi*Exp[1]*x/LambertW[x], 30]
Esta Hoja de cálculo de Excel fórmula utiliza Andre LeClaire la fórmula para aproximar la zeta de Riemann de ceros:
=SI(O(FILA()=1; COLUMNA (A)=1);0; SI(FILA()>=COLUMNA();EXP(-(1-11/8/(COLUMNA (A)-1))/EXP(1)*SUMA(INDIRECTO(DIRECCION(FILA()-COLUMNA()+1; COLUMNA(); 4)&":"&DIRECCIÓN(FILA()-1; COLUMNA(); 4); 4)));0))
(European punto-punto y coma)
es necesario dividir el resultado con: /2/PI()/EXP(1) y tomar el recíproco.
tetration esto es.