Open mezzarobba opened 2 years ago
sage: z = SR.var('z')
sage: expr = 1/(1-z)^(1/2)*log(1/(1-z))^2
sage: ref = list(expr.series(z, 1000).truncate().polynomial(QQ))
sage: asy = asymptotic_expansions.SingularityAnalysis('n', alpha=1/2, beta=2, normalized=False, precision=6)
sage: ini = asy.truncate(5).exact_part().symbolic_expression()
sage: [(RBF(ref[n] - ini(n=n))*RBF(n)**(3/2)) for n in range(980, 1000)]
[[0.640997960 +/- 4.69e-10],
[0.640997394 +/- 2.43e-10],
[0.640996829 +/- 1.62e-10],
[0.640996265 +/- 3.37e-10],
[0.640995702 +/- 4.93e-10],
[0.640995139 +/- 5.73e-10],
[0.640994578 +/- 4.68e-10],
[0.640994018 +/- 3.68e-10],
[0.640993459 +/- 2.95e-10],
[0.640992901 +/- 2.11e-10],
[0.640992344 +/- 1.54e-10],
[0.640991788 +/- 9.93e-11],
[0.640991233 +/- 1.51e-10],
[0.640990679 +/- 2.13e-10],
[0.640990126 +/- 2.71e-10],
[0.640989574 +/- 3.37e-10],
[0.640989023 +/- 4.21e-10],
[0.640988473 +/- 5.19e-10],
[0.640987923 +/- 5.86e-10],
[0.640987375 +/- 4.60e-10]]
oops
I believe that the coefficient of
n^(-3/2)
inis incorrect. Bruno Salvy's
equivalent
(in Maple) confirms and gives c + 11/(3√π) (≈ 0.64) instead, where c ≈ -1.42 is the value in the above output.CC: @cheuberg @behackl @dkrenn
Component: asymptotic expansions
Issue created by migration from https://trac.sagemath.org/ticket/33994