1
$\begingroup$

I have the following recurrence equation: $$(\mu\ n + \nu) f_{n} + J\Phi^{*} \sqrt{n+1}f_{n+1} + J\Phi\ \sqrt{n}f_{n-1} = 0$$ for complex numbers $f_{n}$ where $n = 0,1,2,3,...,\infty$ and complex $\Phi$ and real $\mu, \nu, J$. Is there a way to find a general expression for $f_{n}$ in terms of $f_0$?

A few first terms: $$n = 0:\ \ \ \nu\ f_{0} + J\Phi^{*}f_{1} = 0 \rightarrow f_{1} = -\frac{\nu}{J\Phi^{*}}f_{0}$$ $$n = 1:\ \ \ (\mu + \nu)\ f_{1} + J\Phi^{*}\sqrt{2} f_{2} + J\Phi f_{0} = 0 \rightarrow f_{2} = -\frac{\Phi}{\Phi^{*}\sqrt{2}}f_{0} - \frac{\mu + \nu}{J\sqrt{2}\Phi^{*}}f_{1}$$

$\endgroup$
2
  • $\begingroup$ Since you wrote that this comes from mathematical physics, and out of sheer curiosity, where does this problem come from? $\endgroup$
    – Amir Sagiv
    Apr 14, 2016 at 15:23
  • $\begingroup$ @AmirSagiv This is equation for ground state of the Gutzwiller bosonic wavefunction when you deal with hopping hamiltonian - cold atoms in optical lattice. $\endgroup$
    – WoofDoggy
    Apr 14, 2016 at 19:03

1 Answer 1

2
$\begingroup$

Multiply the equation with $x^n/sqrt(n!)$. Define a generating function $g(x)$ with coefficients $g_n:=f_n/sqrt(n!)$.

You will get a homogeneous differential equation for $g(x)$.

$\endgroup$
1
  • 1
    $\begingroup$ The answer above is an answer to the question, because: It's simple to create the differential equation, which can be easily solved. I don't think, that it is necessary to explain elementary transformations here. @Nex_Friedrich: To develop the result for g(x) into a Taylor series gives the result for g_n and therefore for f_n . $\endgroup$
    – user90369
    Apr 15, 2016 at 7:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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