0

Fix some sequence $(a_k)$ such that $a_k=R_k+q^k$ and $R_k=\frac{1}{R(q^k,q)}$,where $R(x,q)$ is a Generalized Rogers-Ramanujan continued fraction for complex number $x$ and $q=\exp{(2\pi i\tau)}$ is the nome,$|q|\lt1$,and consider $$F(q)=\left(\sum_{n=1}^{\infty} \frac{(-1)^n n}{A_{n-1}}q^{n(n+1)/2}\right)^2- \sum_{n=1}^{\infty} \left(\frac{(-1)^n n}{A_{n-1}}q^{n(n+1)/2}\right)^2\tag1$$ which is satisfied by identity(3) in the other post $$\frac{(q)^5_{\infty}}{(q^5)_{\infty}}=1+5\sum_{n=1}^{\infty} \frac{(-1)^n n}{A_{n-1}} q^{n(n+1)/2}$$ where $$A_n=\prod_{k=1}^{n-1} a_{k}$$ On the other hand, consider, for every positive $k$,$$ G_k(q)=\frac{k}{A_{k-1}}\sum_{n=2}^{\infty} \frac{(-1)^{n-1}}{A_{n+k-2}}(n+k-1)q^{(n^2+(2k-1)n+2k^2)/2}\tag2$$

How to show that $$F(q)=2\sum_{k=1}^\infty G_k(q)\ ?$$

As other commenters suggested,I'll provide a motivating special case(around the boundary of convergence of the Generalized Rogers-Ramanujan continued fraction)

If we consider the fact that $\lim_{x\rightarrow 0+}\frac{1}{R(e^{-nx},e^{-x})}=\phi$ where $\phi=\frac{1+\sqrt5}{2}$ is the golden ratio and applying our conjectured identity,we are led to $$F(1)=\Big(\sum_{n=1}^{\infty}\frac{(-1)^n n}{(1+\phi)^n}\Big)^2-\sum_{n=1}^{\infty}\Big(\frac{(-1)^n n}{(1+\phi)^n}\Big)^2=\frac{1}{25}(1-3\sqrt5)$$

and

$$F(1)=\sum_{k=1}^{\infty}\Big(\sum_{n=2}^{\infty} \frac{(-1)^{n-1}}{(1+\phi)^{n+2k-1}}(2k)(n+k-1)\Big)=\frac{1}{25}(1-3\sqrt5)$$

Nicco
  • 2,853
  • 4
    How do you prove what? What does a series being true mean? Is 1+2+3+4.... true or false? What does that mean? – fleablood Aug 18 '17 at 04:43
  • Please check the inputs as the lhs is $0$ – Supriyo Halder Aug 18 '17 at 06:36
  • 1
    Tried to make this at least readable, but the result is far from perfect. In particular, how are the exponents $\frac{n^2+n+2}{2}$, $\frac{n^2+3n+8}{2}$, etc., computed? – Did Aug 18 '17 at 09:22
  • I think I deciphered the exponents... Now, the lacking pieces are conditions on $(a_n)$ for $F(x)$ and $G(x)$ to be well defined, and, perhaps even more importantly, some context from the OP: either where this identity is coming from, and/or what they tried to prove it. – Did Aug 18 '17 at 09:29
  • @Did :Thanks for your edit.Well I was investigating identity $(3)$ in this post ,and while using mathematica I noticed that the identity proposed in this post related to the infinite series $(3)$ in that post holds.Thus it was deduced from empirical evidence. – Nicco Aug 18 '17 at 09:47
  • 2
    Please make more explicit the alleged relation between the other post and the functions $F$ and $\sum G_k$ here. Please add some conditions on $(a_n)$ under which the identity here holds and interests you. – Did Aug 18 '17 at 09:52
  • I do know a bit about Rogers Ramanujan continued fractions, but I guess your question need far more context than the one available currently so that someone can actually help out. – Paramanand Singh Aug 19 '17 at 07:10
  • Also I don't see the need for $a_{n} $. Your entire questions seems to deal with $A_n$. Further the link to other post is actually a link to this question itself. May be a typo while adding link. – Paramanand Singh Aug 19 '17 at 07:14
  • 1
    Thanks for the corrected link. Btw can you shed some light on the genesis of $G_{k} (q) $. Essentially you are trying to square an infinite series leasing formula $$(\sum a_{i}) ^{2}=\sum a_{i} ^{2}+2\sum_{i<j}a_{i}a_{j}$$ and you have somehow found the expression for $\sum_{i<j} a_{i} a_{j} $ and you want to verify the same here. – Paramanand Singh Aug 20 '17 at 03:12
  • 1
    I don't understand your "motivating example" as it doesn't motivate anything : $\sum_{n=1}^{\infty}\frac{(-1)^n n q^n}{(1+\phi)^n}$ and $\sum_{n=2}^{\infty} \frac{(-1)^{n-1}q^n}{(1+\phi)^{n+2k-1}}(2k)(n+k-1)$ are just rational functions – reuns Aug 20 '17 at 18:46
  • @reuns:What did you mean by "motivating example"? My intention was to provide a special case of the identity around the boundary of convergence(unit circle) of the Generalized Rogers-Ramanujan continued fraction – Nicco Aug 20 '17 at 18:58

0 Answers0