MHB Sturm-Liouville Problem Cheat Sheet

  • Thread starter Thread starter Ackbach
  • Start date Start date
Click For Summary
The discussion provides a comprehensive overview of the Sturm-Liouville problem, detailing the conditions under which the theorem applies, including the necessity for p(x) to be positive and continuous. It establishes that solutions exist only for real values of k under specified boundary conditions, with implications for the completeness and orthogonality of solutions. Modifications to the problem, such as introducing a weighting function, lead to weighted orthogonality instead of standard orthogonality. Several examples illustrate the application of the theorem, showcasing specific cases and their corresponding solutions. This resource is intended to assist students in understanding the Sturm-Liouville framework.
Ackbach
Gold Member
MHB
Messages
4,148
Reaction score
94
This is a helpful document I got from one of my DE's teachers in graduate school, and I've toted it around with me. I will type it up here, as well as attach a pdf you can download.

THEOREM: Consider the Sturm-Liouville problem:
$$p(x) \, y''+p'(x) \, y'+q(x) \, y=k \, y$$
on the interval $[a,b]$ with $p(x)$ and $q(x)$ continuous, and $p(x)>0$. Assume either classical separated boundary conditions, namely, one of:
$$y(x)=0,\qquad y'(x)=0, \qquad y'(x)=c\, y(x)$$
at each endpoint $x=a$ and $x=b$; or else periodic boundary conditions:
$$y(a)=y(b), \qquad y'(a)=y'(b).$$
Then:
  1. the only solutions are for $k$ real.
  2. the set of all linearly independent solutions is complete for the usual
    space of functions.
  3. the set of linearly independent solutions is orthogonal.

COROLLARY: If the condition $p(x)>0$ fails because, at one endpoint, $p(x)=0$, then the classical separated boundary condition at that endpoint is replaced by:
$$y(x)<\infty$$
at that endpoint.

COROLLARY: If the Sturm-Liouville problem is modified by:
$$p(x) \, y''+p'(x) \, y'+q(x) \, y=k \, r(x) \, y$$
with $r(x)$ positive and continuous, and all other conditions the same, then the conclusions of the Theorem are still true, except that the orthogonality conclusion 3. is replaced by a "weighted orthogonality" with weighting function $r(x)$:
$$\int_a^b y_n(x) \, y_m(x) \, r(x) \, dx=0 \quad \text{whenever} \quad n\not=m.$$

Example 1: $y''=ky$ on $[0,\ell]$ with $y(0)=0, \; y(\ell)=0$. Then there are solutions only for $k=-\lambda_n^2=-\dfrac{n^2 \pi^2}{\ell^2}, \; n=1,2,3,\dots$ and $y_n=c_n \, \sin\left(\dfrac{n \pi x}{\ell} \right)$. This gives the Sine series.

Example 2: $y''=ky$ on $[0,\ell]$ with $y'(0)=0, \; y'(\ell)=0$. Then there are solutions for $k=-\lambda_n^2=-\dfrac{n^2 \pi^2}{\ell^2}, \; n=1,2,3,\dots$ and for $k=0$. In the second case, $y=1$ and for $n=1,2,3,\dots, \; y_n=c_n \, \cos\left(\dfrac{n \pi x}{\ell} \right)$. This gives the Cosine series.

Example 3: $y''=ky$ on $[a,b]$ with $y(a)=y(b), \; y'(a)=y'(b)$. Then there are solutions for $k=-\lambda_n^2=-\dfrac{4 n^2 \pi^2}{(b-a)^2}, \; n=1,2,3,\dots$ as well as $k=0$. Further, for each $n\ge 1$ there are two linearly independent solutions, $y_n=a_n \, \sin\left(\dfrac{2n\pi x}{b-a}\right)$ and $y_n=b_n \, \cos\left(\dfrac{2n \pi x}{b-a}\right)$. For $k=0, \; y_0=a_0$. This gives the Fourier series.

Example 4: $xy''+y'=kxy$ with $y(0)<\infty$ and $y(\ell)=0$. There are solutions for certain $k_n=-\lambda_n^2<0, \; n=1,2,3,\dots$ Call the solutions $g_n(x)$. Then the functions $g_n(x)$ are complete on the interval $[0,\ell]$ and are
orthogonal with respect to the weight $x$:
$$\int_0^{\ell}g_m(x) \, g_n(x) \, x \, dx=0 \quad \text{whenever} \quad n\not=m.$$
It turns out that $g_n(x)=J_0(\lambda_n x)$ for the zeroth Bessel function $J_0(x)$.

Attached file: https://www.physicsforums.com/attachments/4725
 

Attachments

Last edited by a moderator:
Physics news on Phys.org
Hope this will be helpful for students. Thanks for sharing!
 

Similar threads

  • · Replies 1 ·
Replies
1
Views
393
  • · Replies 2 ·
Replies
2
Views
5K
  • · Replies 7 ·
Replies
7
Views
4K
  • · Replies 11 ·
Replies
11
Views
4K
  • · Replies 8 ·
Replies
8
Views
5K
Replies
6
Views
2K
  • · Replies 7 ·
Replies
7
Views
3K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 4 ·
Replies
4
Views
2K
  • · Replies 4 ·
Replies
4
Views
2K