MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I am wondering whether there exists a closed form for the definite integral $$F(x)=\int_0^1t^{-a}(1-t)^{N}(1-xt)^{-a}{}_2F_1(-a,k-a-1/2,k-a;4xt(1-xt))dt,$$ where $a\in(0,1)$ and $N,k$ are positive integers.

It seems that I cannot apply Kummer's quadratic transformation formula directly. So I tried to follow the method of proving Kummer's formula by expanding $_2F_1$, and I require some formula for $_2F_1(A,-A,C;x)$. However I cannot find anything relevant in many literatures.

Do you have some other ideas?

Thanks in advance.

share|cite|improve this question

The integral in question is given by,

$$I = \int_{0}^1 t^{-a} (1-t)^N (1-xt)^{-a}F[-a,k-a-1/2,k-a;4xt(1-xt)] \, \mathrm{d}t$$

where $F$ denotes the standard hypergeometric series which is only well-defined for $k-a\geq 0$, and the Pochhammer symbol only terminates if either, $a \geq 0$ or $k-a-1/2 \leq 0.$ Expanding the series yields,

$$F = \sum_{n=0}^\infty \frac{(-a)_{n} (k-a-1/2)_n}{(k-a)_n} \frac{(4xt(1-xt))^n}{n!}$$

where $(q)_n$ denotes the rising Pochhammer symbol (contrary to popular notation). We approach the original integral $I$ by integrating term by term as we slowly expand $F$. The case $n=0$ is,

$$I_{0}=\int_{0}^1 t^{-a}(1-t)^N (1-xt)^{-a} \, \mathrm{d}t$$

as the hypergeometric series is unity for $n=0.$ Notice the integral $I_0$ is precisely an Appell series, i.e.

$$I_0= \frac{\Gamma (1-a)\Gamma(N+1)}{\Gamma(N+a+1)}\mathcal{F}[(1-a),a,0,(N+a+1);x,0]$$

where $\mathcal{F}$ denotes the Appell series, rather than the hypergeometric series. We may proceed similarly for the subsequent $n=1$ contribution, namely, $$I_1 = 4x\frac{a^2(1-a)(k-a-1/2)^2(k-a+1/2)}{(k-a)^2(k-a-1)}\int_0^1 t^{-a+1}(1-t)^N (1-xt)^{-a+1}$$

Luckily $I_1$ is essentially an Appell series also, namely,

$$I_1 = 4x\frac{a^2(1-a)(k-a-1/2)^2(k-a+1/2)}{(k-a)^2(k-a-1)} \frac{\Gamma(2-a)\Gamma(N+1)}{\Gamma(N+3-a)} \\ \times\mathcal{F}[(2-a),(a-1),0,(N+a+1);x,0]$$

Proceed similarly for $n\geq 2$. Goodluck, it's quite a messy problem! Maybe after several iterations you may deduce a general expression for the $n$th contribution $I_n$.

share|cite|improve this answer

May be use the series for $_{2}F_{1}$ and then integrate term by term? The result looks like a sort of Appell-type function of two arguments: $x, 1-x$...

share|cite|improve this answer

The Maple 18 code

A := eval(convert(hypergeom([-a, k-a-1/2], [k-a], y), FPS, y), y = 4*x*t*(-t*x+1)) assuming k::posint:
int(A*t^(-a)*(1-t)^N*(-t*x+1)^(-a), t = 0 .. 1) assuming a>0,a<1,N::posint,x>0,x<1;

outputs the integral under consideration as the series $$\Gamma(N+1)\times$$ $$\sum\limits_{i=0}^\infty \frac{\mathop{\rm pochhammer}(-a,i)x^{i}2^{2i}{\mbox{$_2$F$_1$}(- i+a,1-a+ i;-a+ i+2+N;x)}}{\mathop{\rm pochhammer}(k-a,i)\Gamma(-a+i+2+N)\mathop{\rm pochhammer}(k-a-1/2,i)^{-1}i!},$$ where $\mathop{\rm pochhammer}$ is described here. See here for the Maple output.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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