Wednesday, November 17, 2010

Feigenbaum Constant Approximations

Feigenbaum Constant Approximations
DOWNLOAD Mathematica Notebook

A curious approximation to the Feigenbaum constant delta is given by

 pi+tan^(-1)(e^pi)=4.669201932...,
(1)

where e^pi is Gelfond's constant, which is good to 6 digits to the right of the decimal point.

M. Trott (pers. comm., May 6, 2008) noted

 delta approx 2G+3,
(2)

where G is Gauss's constant, which is good to 4 decimal digits, and

 delta approx 9/T,
(3)

where T is the tetranacci constant, which is good to 3 decimal digits.

A strange approximation good to five digits is given by the solution to

 x^x=1333,
(4)

which is

 x=e^(W(ln1333))=4.669202878...,
(5)

where W(z) is the Lambert W-function (G. Deppe, pers. comm., Feb. 27, 2003).

 delta approx (10)/(pi-1)
(6)

gives delta to 3 digits (S. Plouffe, pers. comm., Apr. 10, 2006).

M. Hudson (pers. comm., Nov. 20, 2004) gave

delta approx (1182102)/(773825)+pi
(7)
 approx (46875)/(15934)-sqrt(2)+pi
(8)
 approx tan((1954)/(1781))+e,
(9)

which are good to 17, 13, and 9 digits respectively.

Stoschek gave the strange approximation

 delta approx 4(1+(12^2)/(163)+(4·12^2+31)/(4·163^2)+...)/(1+(10^2)/(163)+(10^2+30)/(163^2)+...),
(10)

which is good to 9 digits.

R. Phillips (pers. comm., Sept. 14, 2004-Jan. 25, 2005) gave the approximations

delta approx 3/2pi-e^(-pi)
(11)
 approx pi+e-tan^(-1)|alpha|
(12)
 approx (e^(10)-e^9)/(e^8+1)
(13)
 approx 3/2pi-(e^(-pi))/(1+exp(-8+e^(-1/2)))
(14)
 approx pi-tan^(-1)[(e-1)^(-16)-e^pi],
(15)
 approx (e(e-1))/(1+exp{8[(1+e^(-8))^(3/2)-2]})
(16)

where e is the base of the natural logarithm and e^pi is Gelfond's constant, which are good to 3, 3, 5, 7, 9, and 10 decimal digits, respectively, and

|alpha| approx (e/(e-1))^2
(17)
 approx tan(e-delta)
(18)
 approx tan[e-tan^(-1)(e^pi)]
(19)
 approx -cot(e+e^(-pi))
(20)
 approx tan[e+tan^(-1)(2/((e-1)^8e)-e^pi)]
(21)
 approx (e^2)/((e-1)^2-e^(-(3+sqrt(26))))
(22)
 approx (e^2)/((e-1)^2-exp(-8-e^(-1/lnlndelta)),)
(23)

which are good to 3, 3, 3, 4, 6, 8, and 8 decimal digits, respectively.

An approximation to mu_infty due to R. Phillips (pers. comm., Jan. 27, 2005) is obtained by numerically solving

 x=e^(sqrt(phi))(1+2/(e^8lnx)),
(24)

for x, where phi is the golden ratio, which is good to 4 digits.

SEE ALSO: Almost Integer, Feigenbaum Constant