7

This question is a continuation of this post.


Let $r,l,L\geq 1$ be integers. Assume that $q\in [0,1]$ is a real number.

The authors obtained the following equation $36$ in their paper (I express in summation forms).

$$\sum_{r=1}^\infty \sum_{l=1}^\infty \frac{r}{r+l} 4(1-q)q \frac{Γ(L)}{Γ(L − 2p)} \frac{Γ(L + l − 1 − 2p)}{Γ(L + l − 2q)} \frac{Γ(L + l + r − 1 − 2q)}{ Γ(L + l + r) } = $$ $$ 1 - \frac{2q^2}{(1-2q)^2} - \frac{2\pi q(1-q)}{(1-2q)(3-4q)} \cot(2\pi q).$$ Note that the above formula are equal in the sense of asymptotically.

In the midst of obtaining the above formula, I stuck at the following.

Question: Is it possible to express $$\sum_{l=0}^\infty \sum_{r=0}^\infty\frac{\Gamma(L+r-2q)}{\Gamma(L+r-1+2q)} \frac{\Gamma(L+r+l-1+2q)}{\Gamma(L+r+l+2)}\frac{r+1}{r+l+2}$$ in closed form independent of summations (possibly in terms of Gamma function)?

I tried to express above double summations using Appell's series. However, I stuck at $$\frac{\Gamma(L-2q)}{\Gamma(L+2)} \sum_{l=0}^\infty \sum_{r=0}^\infty \frac{(L-1+2q)_{l+r}}{(L+2)_{l+r}} \frac{(L-2q)_r}{(L-1+2q)_r} \frac{(1)_l}{l!r!} \frac{(2)_r}{(r+l+2)}.$$

In particular, can we express $$\frac{(L-2q)_r}{(L-1+2q)_r (r+l+2)}$$ to be something which allows us to use Appell series?


I also tried to evaluate the summation over $l$, that is, $$\sum_{l=0}^\infty \frac{\Gamma(L+r-1+2q+l)}{\Gamma(L+r+2+l)(r+l+2)},$$ and I arrive at $$_3F_2(L+r-1+2q,1,r+2; L+r+2,r+3;1)\frac{\Gamma(L+r-1+2q)}{\Gamma(L+r+2)(r+2)}.$$ However, I could not find any formula to evaluate $_3F_2(L+r-1+2q,1,r+2; L+r+2,r+3;1)$ in Wolfram Alpha.


Updated (2 Nov 2018): @Nikos Bagis uses the Mathematica 10 to obtain a closed form for my question. However, I would like to have a detailed calculations on how to obtain the answer.

Idonknow
  • 15,643

2 Answers2

3

... $$ \sum^{\infty}_{l=0}\sum^{\infty}_{r=0}\frac{\Gamma(L+r-2q)\Gamma(L+r+l-1+2q)}{\Gamma(L+r-1+2q)\Gamma(L+r+l+2)}\frac{r+1}{r+l+2}\tag 0 $$

I have an answer but I use Mathematica 10 for it. Set $$ A_0(L,q,x):=\sum^{\infty}_{l=0}\sum^{\infty}_{r=0}\frac{\Gamma(L+r-2q)\Gamma(L+r+l-1+2q)}{\Gamma(L+r-1+2q)\Gamma(L+r+l+2)}(r+1)x^{l+r+1}. $$ Then using Mathematica evaluate $A_0(L,q,x)$ and FullSimplify it. You will get result. Then integrate the result $$ \int^{1}_{0}A_0(L,q,t)dt $$ and FullSimplify it. You get a function of $\cot$, $\Gamma$, ${}_2F_1$ and ${}_3F_2$ functions. But you have to wait a while.

The function ${}_2F_1$ is computable with elementary functions. It is actualy $$ {}_2F_1(1,-1+L-2q;L;1)=\frac{L-1}{2q} $$ For the ${}_3F_2$, you have to evaluate $$ {}_3F_2(1,1+L-2q,-1+L+2q;2+L,L+2q;1),\tag 1 $$ $$ {}_3F_2(2,2+L-2q,L+2q;3+L,1+L+2q;1)\tag 2 $$ and $$ {}_3F_2(3,3+L-2q,1+L+2q;4+L,2+L+2q;1).\tag 3 $$ Note that these are the zero, first and second derivatives of $$ {}_3F_2(1,1+L-2q,-1+L+2q;2+L,L+2q;x)\tag 4 $$ at $x=1$. So actualy, you have to compute $(4)$, or if you prefere the values $(1),(2),(3)$. The function $(4)$ also referes to the problem of finding $$ \sum^{\infty}_{n=0}\frac{\Gamma(n+a)\Gamma(n+b)}{\Gamma(n+c)\Gamma(n+d)}x^n $$ So the final answer is enter image description here

$\ldots$

CONTINUED

Assume that $$ F_1(L,q,x):=\sum^{\infty}_{l=0}\sum^{\infty}_{r=0}\frac{\Gamma(L+r-2q)\Gamma(L+r+l-1+2q)}{\Gamma(L+r-1+2q)\Gamma(L+r+l+2)}x^lx^r. $$ Then $$ F_1(L,q,x)=-\frac{\Gamma(L-2q)(L-2q-1)}{4q-2}{}_2\widetilde{F}_1\left(1,L-2q;L+2;x\right)+ $$ $$ \frac{(L+2q-2)\Gamma(L-2q)}{4q-2}{}_2\widetilde{F}_1\left(1,L+2q-1;L+2;x\right)- $$ $$ -\frac{\Gamma(L-2q)}{4q-2}{}_2\widetilde{F}_1\left(2,L-2q;L+2;x\right), $$ where $$ {}_p\widetilde{F}_q\left(a_1,a_2,\ldots,a_p;b_1,b_2,\ldots,b_q;z\right):= $$ $$ {}_pF_q\left(a_1,a_2,\ldots,a_p;b_1,b_2,\ldots,b_q;z\right)/(\Gamma(b_1)\Gamma(b_2)\ldots\Gamma(b_q)) $$ is the regularized hypergeometric function.

Set now $$ F_2(L,q,x):=\sum^{\infty}_{l=0}\sum^{\infty}_{r=0}\frac{\Gamma(L+r-2q) \Gamma(L+r+l-1+2q)}{\Gamma(L+r-1+2q) \Gamma(L+r+l+2)}rx^lx^r. $$ Then $$ F_2(L,q,x)= $$ $$ =\frac{x\Gamma(L-2q+1)(2+L-6q-4Lq+8q^2)}{2(2q-1)(4q-3)\Gamma(L+3)}{}_2F_1\left(2,L-2q+1;L+3;x\right)- $$ $$ -\frac{\Gamma(L-2q+1)(L+2q-2)}{2(2q-1)(4q-3)\Gamma(L+2)}{}_2F_1(1,L-2q;L+2;x)+ $$ $$ +\frac{\Gamma(L-2q+1)(L+2q-2)}{2(2q-1)(4q-3)\Gamma(L+2)}{}_2F_1(1,L+2q-1;L+2;x)- $$ $$ -x\frac{\Gamma(L-2q+1)}{(4q-3)\Gamma(L+3)} {}_3F_2\left(2,2,L-2q+1;1,L+3;x\right). $$ But $$ F_3(L,q,x) =\sum^{\infty}_{l=0}\sum^{\infty}_{r=0}\frac{\Gamma(L+r-2q) \Gamma(L+r+l-1+2q)}{\Gamma(L+r-1+2q) \Gamma(L+r+l+2)}(r+1)x^lx^r= $$ $$ =F_1(L,q,x)+F_2(L,q,x) $$ and $$ S=S(L,q)=\int^{1}_{0}x\cdot F_3(L,q,x)dx $$ But also $$ k_{11}(b,c):=\int^{1}_{0}x\cdot { }_2F_1(1,b;c;x)dx= $$ $$ =\frac{(c-1)}{(b-1)(b-2)}\left[2-b+(c-2)\psi(c-2)-(c-2)\psi(c-b)\right], $$ where $Re(b)<Re(c)$. $$ k_{12}(b,c):=\int^{1}_{0}x\cdot {}_2F_1(2,b;c;x)dx= $$ $$ =\frac{(c-1)(c-2)}{(b-1)(b-2)}\left[-1+{}_2F_1(1,b-2;c-2;1)-\psi(c-2)+\psi(c-b)\right] $$ $$ k_{21}(b,c):=\int^{1}_{0}x^2\cdot { }_2F_1(1,b;c;x)dx =\frac{(c-1)}{2(b-1)(b-2)(b-3)}\times $$ $$ \times\left[-(b-3)(2c+b-6)+2(c-3)(c-2)\psi(c-3)-2(c-3)(c-2)\psi(c-b)\right], $$ $$ k_{22}(b,c):=\int^{1}_{0}x^2\cdot { }_2F_1(2,b;c;x)dx =\frac{(c-1)(c-2)}{(b-1)(b-2)(b-3)}\times $$ $$ \times\left[b-c+(c-3){}_2F_1(1,b-3;c-3;1)-2(c-3)\psi(c-3)+2(c-3)\psi(c-b)\right], $$ and $$ M(b,c):=\int^{1}_{0}x^2\cdot { }_3F_2(2,2,b;1,c;x)dx= $$ $$ =-\frac{18}{(b-3)(b-2)(b-1)}-\frac{2b}{(b-3)(b-2)(b-1)}+ $$ $$ \frac{(c-4)(c-3)(c-2)(c-1)}{(b-1)(b-2)(b-c+1)(b-c+2)} +\frac{35c}{(b-1)(b-2)(b-3)}+ $$ $$ +\frac{3bc}{(b-1)(b-2)(b-3)}-\frac{21c^2}{(b-1)(b-2)(b-3)} -\frac{bc^2}{(b-1)(b-2)(b-3)}+ $$ $$ +\frac{4c^3}{(b-1)(b-2)(b-3)}-\frac{4(c-1)(c-2)(c-3)(c-4)}{(b-1)(b-2)(b-3)(c-b-1)}+ $$ $$ +\frac{4(-6+11c-6c^2+c^3)\psi(c-3)}{(b-1)(b-2)(b-3)}-\frac{4(-6+11c-6c^2+c^3)\psi(b-c)}{(b-1)(b-2)(b-3)}, $$ where $2+Re(b)<Re(c)$.

Hence the value of the double sum (tag $(0)$) is $$ S=S(L,q)=-\frac{\Gamma(L-2q)(L-2q-1)}{(4q-2)\Gamma(L+2)} k_{11}(L-2q,L+2)+ $$ $$ +\frac{(L+2q-2)\Gamma(L-2q)}{(4q-2)\Gamma(L+2)}k_{11}(L+2q-1,L+2)- $$ $$ -\frac{\Gamma(L-2q)}{(4q-2)\Gamma(L+2)}k_{12}(L-2q,L+2)+ $$ $$ +\frac{\Gamma(L-2q+1)(2+L-6q-4Lq+8q^2)}{2(2q-1)(4q-3)\Gamma(L+3)}k_{22}(L-2q+1,L+3)- $$ $$ -\frac{\Gamma(L-2q+1)(L+2q-2)}{2(2q-1)(4q-3)\Gamma(L+2)}k_{11}(L-2q,L+2)+ $$ $$ +\frac{\Gamma(L-2q+1)(L+2q-2)}{2(2q-1)(4q-3)\Gamma(L+2)}k_{11}(L+2q-1,L+2)- $$ $$ -\frac{\Gamma(L-2q+1)}{(4q-3)\Gamma(L+3)}M(L-2q+1,L+3). $$ Where $$ \psi(x):=\frac{\Gamma'(x)}{\Gamma(x)}, $$ where $\Gamma(x)$ is the classical Euler's Gamma function.

After simplification we get $$ S=S(L,q)=\frac{(-2 q (q (4 q-11)+8)-2 \pi (q-1) q (2 q-1) \cot (2 \pi q)+3) \Gamma (L-2 q)}{4 (1-2 q)^2 (q-1) q (4 q-3) \Gamma (L)},\tag 5 $$ with the condition $L\in \textbf{R}$ and $0<q<\frac{3}{2}$.

When $q=3/2-1/t$, $t>>1$ and

i) $L=t^a$, $a=1/3$, then $\lim_{t\rightarrow\infty}S(L,q)=1/24$.

ii) $L=t^a$, $a>1/3$, $\lim_{t\rightarrow\infty}S(L,q)=0$

iii) $L=t^a$, $a<1/3$, $\lim_{t\rightarrow\infty}S(L,q)=\infty$

Also for $q\rightarrow 0+$, then if $L\in\{0,-1,-2,\ldots\}$, we get $\lim_{t\rightarrow\infty}S(L,q)=0$.

If $q\rightarrow 0+$ and $L>0$, then $\lim_{q\rightarrow 0+}S(L,q)=\infty$.

  • Possible to include the results shown by Mathematica 10? Otherwise I do not know which functions to evaluate. – Idonknow Nov 02 '18 at 00:20
  • Sorry. What I would like to have is a series of calculations on how to obtain the final answer that you posted. By the way, would you mind to type out the final answer? It is a bit hard to read the image. – Idonknow Nov 02 '18 at 01:20
  • The evaluations are made with Mathematica. I don't know how ecxtracted. But I will try to answer your questions. – Nikos Bagis Nov 02 '18 at 08:16
  • Nikos bagis yeah it would be better if we know how to obtain the answer. – Idonknow Nov 02 '18 at 23:02
  • Sharp the idea of integrating (+1). For the process to obtain the result, probably that is based on the Gosper-Zeilberg algorithm. – G Cab Nov 02 '18 at 23:40
  • @NikosBagis G Cab has expressed the double summations using Kampé de Fériet function. Maybe your answer can be derived from there? – Idonknow Nov 05 '18 at 03:32
  • @NikosBagis possible to use Mathematica 10 to calculate Kampe de Feriet series obtained by @G Cab? Thank. – Idonknow Nov 06 '18 at 14:48
  • @NikosBagis can you elaborate on how to obtain second equation after continue? The one with regular hypergeometric series. – Idonknow Nov 07 '18 at 23:11
  • @NikosBagis In the paper, they stated that if we take limit $q\to 0,$ then the answer should be $\frac{2}{3}.$ However, Wolfram alpha tells us that the limit does not exist. What is wrong? – Idonknow Nov 08 '18 at 07:58
  • @NikosBagis I input your last expression into Wolfram alpha and compute limit when $q\to 0^+.$ Wolfram gave me some werid things. https://www.wolframalpha.com/input/?i=lim_%7Bq%5Cto+0%7D+%5CGamma+(L-2+q)%5Cfrac%7B%5B-2+q+(q+(4+q-11)%2B8)-2+%5Cpi++(q-1)+q+(2+q-1)+%5Ccot+(2+%5Cpi++q)%2B3%5D+%7D%7B4+(1-2+q)%5E2+(q-1)+q+(4+q-3)+%5CGamma+(L)%7D – Idonknow Nov 08 '18 at 12:49
  • @NikosBagis However, in the context, $L$ is a positive integer.... – Idonknow Nov 08 '18 at 14:59
3

The sum to find can be written in the two equivalent forms, as demonstrated in the precedent post. $$ \eqalign{ & S(L,q) = \sum\limits_{r = 0}^\infty {\sum\limits_{l = 0}^\infty {{{\Gamma (L + r - 2q)} \over {\Gamma (L + r - 1 + 2q)}}{{\Gamma (L + r + l - 1 + 2q)} \over {\Gamma (L + r + l + 2)}}{{r + 1} \over {r + l + 2}}} } = \cr & = \sum\limits_{r = 0}^\infty {\sum\limits_{l = 0}^\infty {\left( {r + 1} \right)\left( {L + r - 1 + 2q} \right)^{\,\overline {\, - \left( { 4q-1} \right)\,} } \left( {L + r + l + 2} \right)^{\,\overline {\, - 3 + 2q\,} } {1 \over {r + l + 2}}} } \cr & = \sum\limits_{s = 0}^\infty {{1 \over {s + 2}}\left( {L + s + 2} \right)^{\,\overline {\, - 3 + 2q\,} } \sum\limits_{r = 0}^s {\left( {r + 1} \right)\left( {L + r - 1 + 2q} \right)^{\,\overline {\, - \left( { 4q-1} \right)\,} } } } \cr} $$

Now, using the composition formula for the Rising Factorial $$ x^{\,\overline {\,a + b\,} } = x^{\,\overline {\,a\,} } \left( {x + a} \right)^{\,\overline {\,b\,} } $$ we obtain two fundamental blocks that we need for proceeding $$ \eqalign{ & \left( {r + 1} \right)\left( {r + 1 + a} \right)^{\,\overline {\,1 - b\,} } = {{\left( {r + 1} \right)^{\,\overline {\,1\,} } \left( {r + 2} \right)^{\,\overline {\,a - 1\,} } \left( {r + 1 + a} \right)^{\,\overline {\,1 - b\,} } } \over {\left( {r + 2} \right)^{\,\overline {\,a - 1\,} } }} = \cr & = {{\left( {r + 1} \right)^{\,\overline {\,a + 1 - b\,} } } \over {\left( {r + 2} \right)^{\,\overline {\,a - 1\,} } }} = {{1^{\,\overline {\,r + 1\,} } } \over {1^{\,\overline {\,r\,} } }}{{1^{\,\overline {\,r\,} } \left( {r + 1} \right)^{\,\overline {\,a + 1 - b\,} } } \over {1^{\,\overline {\,r + 1\,} } \left( {r + 2} \right)^{\,\overline {\,a - 1\,} } }} = {{2^{\,\overline {\,r\,} } } \over {1^{\,\overline {\,r\,} } }}{{1^{\,\overline {\,r + a + 1 - b\,} } } \over {1^{\,\overline {\,r + a\,} } }} = \cr & = {{2^{\,\overline {\,r\,} } } \over {1^{\,\overline {\,r\,} } }}{{1^{\,\overline {\,a + 1 - b\,} } \left( {a + 2 - b} \right)^{\,\overline {\,r\,} } } \over {1^{\,\overline {\,a\,} } \left( {a + 1} \right)^{\,\overline {\,r\,} } }} = \left( {a + 1} \right)^{\,\overline {\,1 - b\,} } {{2^{\,\overline {\,r\,} } \left( {a + 2 - b} \right)^{\,\overline {\,r\,} } } \over {1^{\,\overline {\,r\,} } \left( {a + 1} \right)^{\,\overline {\,r\,} } }} \cr} $$ and $$ \eqalign{ & {1 \over {y + 2}}\left( {y + 2 + c} \right)^{\,\overline {\,d\,} } = {{\left( {y + 3} \right)^{\,\overline {\,c - 1\,} } \left( {y + 2 + c} \right)^{\,\overline {\, - d\,} } } \over {\left( {y + 2} \right)^{\,\overline {\,1\,} } \left( {y + 3} \right)^{\,\overline {\,c - 1\,} } }} = \cr & = {{\left( {y + 3} \right)^{\,\overline {\,c - d - 1\,} } } \over {\left( {y + 2} \right)^{\,\overline {\,c\,} } }} = {{1^{\,\overline {\,y + 1\,} } } \over {1^{\,\overline {\,y + 2\,} } }}{{1^{\,\overline {\,y + 2\,} } \left( {y + 3} \right)^{\,\overline {\,c - d - 1\,} } } \over {1^{\,\overline {\,y + 1\,} } \left( {y + 2} \right)^{\,\overline {\,c\,} } }} = \cr & = {{1^{\,\overline {\,y + 1\,} } } \over {1^{\,\overline {\,y + 2\,} } }}{{1^{\,\overline {\,y + 1 + c - d\,} } } \over {1^{\,\overline {\,y + 1 + c\,} } }} = {{1^{\,\overline {\,1\,} } 2^{\,\overline {\,y\,} } } \over {1^{\,\overline {\,2\,} } 3^{\,\overline {\,y\,} } }} {{1^{\,\overline {\,1 + c - d\,} } \left( {2 + c - d} \right)^{\,\overline {\,y\,} } } \over {1^{\,\overline {\,1 + c\,} } \left( {2 + c} \right)^{\,\overline {\,y\,} } }} = \cr & = {{1^{\,\overline {\,1\,} } 1^{\,\overline {\,1 + c - d\,} } } \over {1^{\,\overline {\,2\,} } 1^{\,\overline {\,1 + c\,} } }} {{2^{\,\overline {\,y\,} } \left( {2 + c - d} \right)^{\,\overline {\,y\,} } } \over {3^{\,\overline {\,y\,} } \left( {2 + c} \right)^{\,\overline {\,y\,} } }} = \cr & = {{\left( {c + 2} \right)^{\,\overline {\, - d\,} } } \over 2}{{2^{\,\overline {\,y\,} } \left( {2 + c - d} \right)^{\,\overline {\,y\,} } } \over {3^{\,\overline {\,y\,} } \left( {2 + c} \right)^{\,\overline {\,y\,} } }} \cr} $$

1) the $(r,l)$ version

Applying the above to the sum in $(r,l)$ $$ \bbox[lightyellow] { \eqalign{ & \left\{ \matrix{ \matrix{ {a = L - 2 + 2q} \hfill & {b = \,4q} \hfill \cr {c = L} \hfill & {d = 3 - 2q\,} \hfill \cr } \hfill \cr A = {1 \over 2}\left( {c + 2} \right)^{\,\overline {\, - d\,} } \left( {a + 1} \right)^{\,\overline {\,1 - b\,} } = {1 \over 2}\left( {L + 2} \right)^{\,\overline {\, - 2 - 2q\,} } \hfill \cr} \right. \cr & S(L,q) = \sum\limits_{r = 0}^\infty {\sum\limits_{l = 0}^\infty {\left( {r + 1} \right)\left( {r + 1 + L - 2 + 2q} \right)^{\,\overline {\,1 - 4q\,} } \left( {L + r + l + 2} \right)^{\,\overline {\, - \left( {3 - 2q} \right)\,} } {1 \over {r + l + 2}}} } = \cr & = A\;\sum\limits_{r = 0}^\infty {\sum\limits_{l = 0}^\infty {{{2^{\,\overline {\,r + l\,} } \left( {2 + c - d} \right)^{\,\overline {\,r + l\,} } } \over {3^{\,\overline {\,r + l\,} } \left( {2 + c} \right)^{\,\overline {\,r + l\,} } }}{{2^{\,\overline {\,r\,} } \left( {a + 2 - b} \right)^{\,\overline {\,r\,} } } \over {1^{\,\overline {\,r\,} } \left( {a + 1} \right)^{\,\overline {\,r\,} } }}} } \cr} } \tag{1.1}$$

Now, the above, once rewritten as $$ \bbox[lightyellow] { \eqalign{ & S(L,q) = A\;\sum\limits_{r = 0}^\infty {\sum\limits_{l = 0}^\infty {{{2^{\,\overline {\,r + l\,} } \left( {2 + c - d} \right)^{\,\overline {\,r + l\,} } } \over {3^{\,\overline {\,r + l\,} } \left( {2 + c} \right)^{\,\overline {\,r + l\,} } }} {{2^{\,\overline {\,r\,} } 1^{\,\overline {\,l\,} } \left( {a + 2 - b} \right)^{\,\overline {\,r\,} } 1^{\,\overline {\,l\,} } } \over {\left( {a + 1} \right)^{\,\overline {\,r\,} } 1^{\,\overline {\,l\,} } }}} } {{x^{\,r} } \over {r!}}{{y^{\,l} } \over {l!}} = \cr & = A\,F\left( {\matrix{ 2 \cr 3 \cr 2 \cr 2 \cr } \,\left| {\,\matrix{ {2;\left( {2 + c - d} \right)} \cr {2,\,1\;;\;\left( {a + 2 - b} \right),1\;} \cr {3;\left( {2 + c} \right)} \cr {\;\left( {a + 1} \right),1} \cr } \,} \right|x,y} \right)\quad \quad \left| {\;x = y = 1} \right. \cr} } \tag{1.2}$$

tells us that the sum is a

Kampé de Fériet function computed at $x=y=1$.

Looking on the web one can find various specialized papers dealing with the expansion of Kampé de Fériet function function in terms of Hypergeometric functions. So the expansion found by @Nikos Bagis well enters into the scene. (re. to e.g. this paper and to this and this).

Also interesting is the connection with the "Extended Beta" integral representation (re. for instance to this and this paper, and specially this), which relates with the integral representation hinted in my answer to the previous post

2) the $(r,s)$ truncated sum version

Restarting from (1.1) above, and writing $S(L,q)$ in the truncated way we get $$ \bbox[lightyellow] { \eqalign{ & S(L,q) = \sum\limits_{r = 0}^\infty {\sum\limits_{l = 0}^\infty {\left( {r + 1} \right)\left( {r + 1 + L - 2 + 2q} \right)^{\,\overline {\,1 - 4q\,} } \left( {L + r + l + 2} \right)^{\,\overline {\, - \left( {3 - 2q} \right)\,} } {1 \over {r + l + 2}}} } = \cr & = A\;\sum\limits_{r = 0}^\infty {\sum\limits_{l = 0}^\infty {{{2^{\,\overline {\,r + l\,} } \left( {2 + c - d} \right)^{\,\overline {\,r + l\,} } } \over {3^{\,\overline {\,r + l\,} } \left( {2 + c} \right)^{\,\overline {\,r + l\,} } }}{{2^{\,\overline {\,r\,} } \left( {a + 2 - b} \right)^{\,\overline {\,r\,} } } \over {1^{\,\overline {\,r\,} } \left( {a + 1} \right)^{\,\overline {\,r\,} } }}} } = \cr & = A\;\sum\limits_{s = r + l = 0}^\infty {{{2^{\,\overline {\,s\,} } \left( {2 + c - d} \right)^{\,\overline {\,s\,} } } \over {3^{\,\overline {\,s\,} } \left( {2 + c} \right)^{\,\overline {\,s\,} } }}\sum\limits_{r = 0}^s {{{2^{\,\overline {\,r\,} } \left( {a + 2 - b} \right)^{\,\overline {\,r\,} } } \over {1^{\,\overline {\,r\,} } \left( {a + 1} \right)^{\,\overline {\,r\,} } }}} } \cr} } \tag{2.1}$$

In order to reduce it, exactly or asymptotically, into more manageable terms, I am currently exploring the possibilities offered by some possible approaches that I could individuate up to now.

2.a) Product of two Hypergeometric A truncated Hypergeometric series can be transformed into the limit of a standard one, by adding two additional parameters $$ \eqalign{ & \sum\limits_{r = 0}^s {{{2^{\,\overline {\,r\,} } \left( {a + 2 - b} \right)^{\,\overline {\,r\,} } } \over {1^{\,\overline {\,r\,} } \left( {a + 1} \right)^{\,\overline {\,r\,} } }}} = \mathop {\lim }\limits_{\varepsilon \, \to \,0} \sum\limits_{r = 0}^\infty {{{\left( { - s} \right)^{\,\overline {\,r\,} } 2^{\,\overline {\,r\,} } \left( {a + 2 - b} \right)^{\,\overline {\,r\,} } } \over {\left( { - s + \varepsilon } \right)^{\,\overline {\,r\,} } 1^{\,\overline {\,r\,} } \left( {a + 1} \right)^{\,\overline {\,r\,} } }}} = \cr & = \mathop {\lim }\limits_{\varepsilon \, \to \,0} \sum\limits_{r = 0}^\infty {{{\left( { - s} \right)^{\,\overline {\,\varepsilon \,} } 2^{\,\overline {\,r\,} } \left( {a + 2 - b} \right)^{\,\overline {\,r\,} } } \over {\left( { - s + r} \right)^{\,\overline {\,\varepsilon \,} } 1^{\,\overline {\,r\,} } \left( {a + 1} \right)^{\,\overline {\,r\,} } }}} \cr} $$ (re. to the renowned "Concrete Mathematics" - Ch. 5 - ex. 34)
which takes advantage of that $\left( { - s} \right)^{\,\overline {\,r\,} }$ becomes null for non negative integers $s < r$.
This gives us a summand product of two hypergeometric terms, but with one fraction containing both parameters.
Since $$ {{\left( { - s} \right)^{\,\overline {\,\varepsilon \,} } } \over {\left( { - s + r} \right)^{\,\overline {\,\varepsilon \,} } }} = {{\Gamma \left( { - s + \varepsilon } \right)\Gamma \left( { - s + r} \right)} \over {\Gamma \left( { - s} \right)\Gamma \left( { - s + r + \varepsilon } \right)}} = {{{\rm B}\left( { - s + r,\varepsilon } \right)} \over {{\rm B}\left( { - s,\varepsilon } \right)}} $$ there might be a possibility to expand it in some interesting way.

2.b) series expression of the truncated Hypergeometric

Trying to apply the formulas provided in this work

G Cab
  • 35,272
  • This is good to know. Thanks. – Idonknow Nov 03 '18 at 05:45
  • 1
    However, how does your hint relate in finding a closed form of my question? – Idonknow Nov 03 '18 at 14:42
  • It would be better if you can include an explanation on how to use your hints... – Idonknow Nov 04 '18 at 06:03
  • 1
    @Idonknow: expanded the hint ... found the relation – G Cab Nov 04 '18 at 14:49
  • @G Cab Regarding the interesting connection with 'extended Beta' integral representation, do you mind to elaborate on it? Thanks. – Idonknow Nov 04 '18 at 15:57
  • @G Cab By computing Kampé de Fériet function at $x=y=1,$ we can obtain answer by @Nikos Bagis? I do not see how. Can you elaborate? – Idonknow Nov 04 '18 at 16:02
  • @G Cab By the way, I just did some googling but not able to find any paper connecting expansion of Kampé de Fériet function with Hypergeometric series. Do you mind yo provide at least one reference? – Idonknow Nov 04 '18 at 16:13
  • @Idonknow, Mh-hmm! I cannot compete with M.10 in "dis-intricating " such a formula, nor I have it available. I would ask N.Bagis to go in a step-by-step simplification, in order to possibly catch M.10's process. However, on my side, I'll try with my best "human" math tools to further simplify the above. Pls. consider that you just led me to discover K. de F. function, and its many applications in physics! – G Cab Nov 04 '18 at 17:47
  • @Idonknow: added some more references: but pls. consider that I am quite new to 2-D HyperG. – G Cab Nov 04 '18 at 22:42
  • @G Cab I apologize if I offended you due to my rudeness on your hard work. I agree that discovering K. de F. function indeed does express my double summations into closed form. However, I just want to know further what can be done. – Idonknow Nov 04 '18 at 22:45
  • @Idonknow: absolutely no offence. The matter is new to me but very interesting. I'm trying other ways to express the result. – G Cab Nov 05 '18 at 01:50
  • @G Cab I think there is a typo in $S(L,q)$ before the RIsing factorial. The power of the factor $(L+r-1+2q)$ should be ${\overline{1-4q}}$ instead $\overline{-(1+4q)}.$ – Idonknow Nov 06 '18 at 03:31
  • @G Cab Do identities in section $3.2$ of the paper help? https://www.mat.univie.ac.at/~schlosse/pdf/appell-survey.pdf – Idonknow Nov 06 '18 at 04:28
  • @G Cab Regarding newly added (2.a), may I know how does it help in evaluating the double summations? – Idonknow Nov 06 '18 at 23:27
  • @Idonknow: yes, you are right, there is a typo: sorry I will amend asap – G Cab Nov 08 '18 at 02:13
  • @Idonknow: typos amended, formulas cleansed up and checked, note the change in $A,a,b,c,d$ to render $d$ non-negative. – G Cab Nov 08 '18 at 12:01
  • @GCab Would you happen to know the answer to this question about the reduced sextic and Kampé de Fériet function? – Tito Piezas III Jun 30 '23 at 15:19