1
$\begingroup$

I just tried in Mathematica 11.3

SeriesCoefficient[Hypergeometric2F1[1, 1 - n, 2 + n, w], {w, 0, m}, Assumptions -> n > m > 1 && n \[Element] Integers]

which gives $$ \frac{(n+1)! (m-n)!}{(-n)! (m+n+1)!} $$

This cannot be correct, since $m-n<0$ and its factorial is infinity. Is there any walkaround for this problem?

$\endgroup$
  • 1
    $\begingroup$ You have negative factorials in the numerator and denominator, so the ratio is regular and equal to $\frac{(-1)^m (n-1)! (n+1)!}{(n-m-1)! (n+m+1)!}$ for $n > m > 1$ and $n$,$m$ integers. $\endgroup$ – QuantumDot Jan 13 at 15:25
3
$\begingroup$

I'd say it was a bug. Here's a better expression:

FunctionExpand@
  SeriesCoefficient[Hypergeometric2F1[1, 1 - n, 2 + n, w], {w, 0, m}, 
   Assumptions -> n > m > 1 && n ∈ Integers] //. {Gamma[a_]/
   Gamma[b_] :> Pochhammer[b, a - b]}
(*  Pochhammer[1 - n, m] Pochhammer[2 + m + n, -m]  *)
$\endgroup$
0
$\begingroup$

Here is another approach. First, use the Pfaff transformation on your expression:

f = Hypergeometric2F1[1, 1 - n, 2 + n, w] /. 
    Hypergeometric2F1[a_, b_, c_, z_] :> (1 - z)^(c - a - b) Hypergeometric2F1[c - a, c - b, c, z];

You can then apply SeriesCoefficient[]:

Assuming[m >= 0, SeriesCoefficient[f, {w, 0, m}]]
   (Gamma[m - 2 n] HypergeometricPFQ[{-m, 1 + n, 1 + 2 n}, {2 + n, 1 - m + 2 n}, 1])/
   (Gamma[1 + m] Gamma[-2 n])

Altho a ${}_3 F_2$ function has appeared, this is in a form where the Saalschütz formula can be applied:

% /. HypergeometricPFQ[{a_, b_, c_}, {d_, e_}, 1] :> 
     Pochhammer[d - b, -a] Pochhammer[d - c, -a]/(Pochhammer[d, -a] Pochhammer[d - b - c, -a])
   (Gamma[m - 2 n] Pochhammer[1, m] Pochhammer[1 - n, m])/
   (Gamma[1 + m] Gamma[-2 n] Pochhammer[-2 n, m] Pochhammer[2 + n, m])

FullSimplify[%]
   (Gamma[1 + m - n] Gamma[2 + n])/(Gamma[1 - n] Gamma[2 + m + n])

which is equivalent to the expression Michael obtained, but can also be rewritten as Pochhammer[1 - n, m]/Pochhammer[n + 2, m]. Note that this is consistent with the choice of parameters in the original hypergeometric function, since we have the identity SeriesCoefficient[Hypergeometric2F1[a, b, c, w], {w, 0, m}] == (Pochhammer[a, m] Pochhammer[b, m])/(m! Pochhammer[c, m]).

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.