0
$\begingroup$

The operator $E$ is defined as $Eu_n=u_{n+1}$.

I encountered a strange equality. when I tried out Let $u_n$ represent a series such that

$$u_{n+2}=u_{n+1}+u_n. \tag{$\star$}$$

Or $E^2{u_n}=Eu_{n+1}+Eu_n$

$\left(E^2-E-1\right)u_n=0$

$E\equiv \frac{1\pm \sqrt{5}}{2} $

What interpretation should be made on such an equality ?.

This equality seems interesting to me as two well know sequences that obeys$\huge\star$ Fibonacci sequence($u_1=u_2=1$) and Lucas sequence($u_1=1,u_2=3$) has it's $n$ th term represented by

$\Large {\frac{\left(\frac{1+\sqrt{5}}{2}\right)^n-\left(\frac{1-\sqrt{5}}{2}\right)^n}{\sqrt{5}}}$ and $\Large {\frac{\left(\frac{1+\sqrt{5}}{2}\right)^n+\left(\frac{1-\sqrt{5}}{2}\right)^n}{2}}$ respectively

If I am to follow that $E$ and $\frac{1\pm \sqrt{5}}{2} $. Then in case of the above two particular case of the recursion . The formula for the $n$ the term is not true(Of course I can't determine which sign($\pm$) should be taken and hence such an argument is bad).But interestingly , when we assume the same value for $E$, $\huge \star$ is valid with both the signs(the $\pm$).

I am really interested with such an equivalence, as I think it's interpretation can be exploited for a proof of those formulas. And also possibly a general solution. As

$E^n{u_n}=u_{n+1}$

$\endgroup$

1 Answer 1

4
$\begingroup$

$E$ acts on the finite-dimensional vector space consisting of sequences satisfying that recurrence relation, and those are its eigenvalues on that space. More generally, $E$ acts on the nullspace of $p(E)$ for any polynomial $p$ with eigenvalues the roots of $p$, and you can use this together with the theory of Jordan normal form to write down general solutions to linear recurrence relations.

$\endgroup$

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.