Problems with xrefs to A127585
Jonathan Post
jvospost3 at gmail.com
Wed Apr 4 18:10:28 CEST 2007
Thank you Frank.
Now I can, unless you object, submit a sequence correctly
crossreferencing my A127585, along these lines:
Error term in Stirling's approximation truncated after e^(-n).
a(n) = Floor[(sqrt(2*pi))*(n^(n+(1/2)))*(e^(-n))]
offset 1,5
0, 0, 0, 0, 1, 9, 59, 417, 3343, 30104, 301174, 3314113, 39781324
a(14) ~ 517289459 but these were crudely calculated with the Google
calculator. Before I submit, Frank, are these correct, as done with a
higher precision program?
example:
a(7) = Floor[(sqrt(2*pi))*(7^(7+(1/2)))*(e^(-7))]
= Floor[59.6041684] = 59.
Citation to equation (17) in Weisstein reference "Stirling's
approximation" in MathWorld.
Then the next sequence, which would reference the above and A127585:
Error term in Stirling's approximation truncated after e^(-n).
a(n) = Floor[(sqrt(2*pi))*(n^(n+(1/2)))*(e^(-n))*(e^(1/((12*n)+1)))]
I looked at the Herbert Robbins citation, "A remark of Stirling's
Formula", Amer. Math. Monthly 62 (1955) 26-29, which is a really
lovely little elementary proof. But I don't know the next truncation
from Striling's Series, beyond the double inequality of Robbins(1958)
and Feller (1968).
Does anyone have a citation in Gosper's cute refinement as equation
(18) in the above:
n! ~ sqrt(((2*n) + (1/3))*pi) * (n^n) * (e^(-n))
More information about the SeqFan
mailing list