Regular Sturm-Liouville boundary-value problem on the interval [-4,4] given by $$y''+\lambda y=0$$ with $y(-4)=0$ and $y(4)=0$. Given that $\lambda >0$ the general solution is $$ a\cos\left(\sqrt{\lambda }x\right)+ b\sin\left(\sqrt{\lambda }x\right)$$

have to find all positive eigenvalues of the problem along with the corresponding eigenfunctions. I am confused because when trying to find the eigenvalues I don't have any cancellation, (for example if it said y(0)=0 instead of y(-4)=0.

any help appreciated!!

  • $\begingroup$ Note that $\cos(-4\sqrt{\lambda}) = \cos(4\sqrt{\lambda})$ and $\sin(-4\sqrt{\lambda}) = -\sin(4\sqrt{\lambda})$. Use that to transform the boundary condition for $x=-4$ to an equation similar to the boundary condition for $x=4$. It should then be easy to figure out what the possibilities are for $\lambda$. $\endgroup$ – rogerl Jun 2 '16 at 21:36

There's a nice technique you can use in such problems. Because you're only interested in solutions that are not identically $0$, you can pick one of the endpoints such as $-4$ and add another condition $y(-4)=0$, $y'(-4)=1$. You can assume this because $y'(-4)\ne 0$--otherwise, $y\equiv 0$ due to the uniqueness of solutions of $y''+\lambda y=0$ with $y(-4)=y'(-4)=0$.

The solution of $y''+\lambda y=0$ with $y(-4)=0$, $y'(-4)=1$ is $$ y_{\lambda}(x) = \frac{\sin(\sqrt{\lambda}(x+4))}{\sqrt{\lambda}}. $$ The other nice part of doing things this way is that special cases are handled by limiting cases. This is because you are guaranteed that two fixed endpoint conditions will result in a solution that has an everywhere convergent power series expansion in $\lambda$. That's a general theorem of ODEs. So, $$ y_{0}(x) = \lim_{\lambda\rightarrow 0}y_{\lambda}(x)=x+4. $$ You can check--but you don't have to--that this is solution where $\lambda=0$.

Any non-zero solution of $y''+\lambda y=0$ with $y(-4)=0$, $y(4)=0$ must be a non-zero multiple of $y_{\lambda}(x)$. Therefore, the permissible eigenvalues $\lambda$ are the solutions of $$ \frac{\sin(\sqrt{\lambda}8)}{\sqrt{\lambda}}=0. $$ Checking separately, $y_0$ does not work. The only $\lambda$ that work are $$ \sqrt{\lambda}8 = \pm \pi,\pm 2\pi,\pm 4\pi,\cdots, \\ \lambda = \frac{n^2\pi^2}{64},\;\; n=1,2,3,\cdots. $$


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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