Difference between revisions of "Kneser expansion"

From TORI
Jump to navigation Jump to search
 
 
(One intermediate revision by the same user not shown)
Line 3: Line 3:
 
While in this article, only the special natural superexponential is considered. In this case, the fixed point is
 
While in this article, only the special natural superexponential is considered. In this case, the fixed point is
   
  +
\(
$
 
L= (-\mathrm{ProductLog}(-1))^*\approx$ $
+
L= (-\mathrm{ProductLog}(-1))^*\approx\) \(
0.3181315$ $+$
+
0.3181315\) \(+\)
$1.3372357 \,\mathrm i
+
\(1.3372357 \,\mathrm i
  +
\)
$
 
   
The solution of $F$ of the [[transfer equation]]
+
The solution of \(F\) of the [[transfer equation]]
   
  +
\(
$
 
 
F(z+1)=\exp(F(z))
 
F(z+1)=\exp(F(z))
  +
\)
$
 
   
 
is asymptotically expanded in the gollowing form:
 
is asymptotically expanded in the gollowing form:
   
$ \displaystyle
+
\( \displaystyle
 
F(z)=L+\exp(L z)+\sum_{n=2}^{M_0} a_{0,n} \exp(L n z) + O(\exp(L (M_0\!+
 
F(z)=L+\exp(L z)+\sum_{n=2}^{M_0} a_{0,n} \exp(L n z) + O(\exp(L (M_0\!+
\!1)z)$
+
\!1)z)\)
$\displaystyle + b_1 \exp(2 \pi \mathrm i z) \left(1+\sum_{n=1}^{M_1} a_{1,n} \exp(L n z) + ..\right)$
+
\(\displaystyle + b_1 \exp(2 \pi \mathrm i z) \sum_{n=1}^{M_1} a_{1,n} \exp(L n z) + ..\)
$\displaystyle + b_2 \exp(4 \pi \mathrm i z) \left(1+\sum_{n=1}^{M_2} a_{2,n} \exp(L n z) + ..\right) + .. $
+
\(\displaystyle + b_2 \exp(4 \pi \mathrm i z) \sum_{n=1}^{M_2} a_{2,n} \exp(L n z) + .. \)
   
  +
..
   
Coefficients $a_{0,n}$ for $n=1..12$ can be evaluated with the [[Mathematica]] code below
+
Coefficients \(a_{0,n}\) for \(n=1..12\) can be evaluated with the [[Mathematica]] code below
 
<poem><nomathjax><nowiki>
 
<poem><nomathjax><nowiki>
 
M=12; f0 = L + e + Sum[a[n] e^n, {n,2,M}];
 
M=12; f0 = L + e + Sum[a[n] e^n, {n,2,M}];

Latest revision as of 22:50, 15 August 2020

Kneser expansion is asymptotic representation of superexponential constructed at its fixed point.

While in this article, only the special natural superexponential is considered. In this case, the fixed point is

\( L= (-\mathrm{ProductLog}(-1))^*\approx\) \( 0.3181315\) \(+\) \(1.3372357 \,\mathrm i \)

The solution of \(F\) of the transfer equation

\( F(z+1)=\exp(F(z)) \)

is asymptotically expanded in the gollowing form:

\( \displaystyle F(z)=L+\exp(L z)+\sum_{n=2}^{M_0} a_{0,n} \exp(L n z) + O(\exp(L (M_0\!+ \!1)z)\) \(\displaystyle + b_1 \exp(2 \pi \mathrm i z) \sum_{n=1}^{M_1} a_{1,n} \exp(L n z) + ..\) \(\displaystyle + b_2 \exp(4 \pi \mathrm i z) \sum_{n=1}^{M_2} a_{2,n} \exp(L n z) + .. \)

..

Coefficients \(a_{0,n}\) for \(n=1..12\) can be evaluated with the Mathematica code below


M=12; f0 = L + e + Sum[a[n] e^n, {n,2,M}];
f1 = L + L e + Sum[a[n]L^n e^n, {n,2,M}];
s0 = ReplaceAll[Series[Exp[f0],{e,0,M}] - f1, Exp[L]->L];
co[2] = Extract[Solve[Coefficient[s0,e^2]==0, a[2]], 1];
A[2] = ReplaceAll[a[2], co[2]]
s[3] = Simplify[ReplaceAll[s0, a[2] -> A[2]]];

For[m = 3, m <= M, Print[m];
co[m] = Extract[Solve[Coefficient[s[m], e^m] == 0, a[m]], 1];
A[m] = ReplaceAll[a[m], co[m]];
s[m+1] = Simplify[ReplaceAll[s[m], a[m]->A[m]]]; m++]

tableM = Table[{a[m], A[m]}, {m,2,M}];
Le = N[Conjugate[-ProductLog[-1]], 64]
N[TableForm[Table[{a[m], ReplaceAll[A[m], L -> Le]}, {m, 2, M}]], 16]

References

http://www.digizeitschriften.de/dms/img/?PPN=GDZPPN002175851 H.Kneser. Reelle analytische Lösungen der Gleichung φ(φ(x))=ex. Equationes Mathematicae, Journal fur die reine und angewandte Mathematik {\bf 187} 56–67 (1950)

http://mizugadro.mydns.jp/PAPERS/2010vladie.pdf D.Kouznetsov. Superexponential as special function. Vladikavkaz Mathematical Journal, 2010, v.12, issue 2, p.31-45.

Keywords

Iteration, Tetration, Transfer equation, Transfer function, Superexponential Superfunction,,,,