Sturm-Liouville theory/Proofs: Difference between revisions

From Citizendium
Jump to navigation Jump to search
imported>Dan Nessett
(→‎Proof: fixed links so they point to CZ rather than WP article)
imported>Dan Nessett
(Adjust vertical alignment on some equations)
 
(5 intermediate revisions by the same user not shown)
Line 1: Line 1:
{{subpages}}
{{subpages}}
__TOC__


This article proves that solutions to the Sturm-Liouville equation corresponding to distinct eigenvalues are orthogonal. Note that when the Sturm-Liouville problem is regular, distinct eigenvalues are guaranteed. For background see [[Sturm-Liouville theory]].
This article proves that solutions to the Sturm-Liouville equation corresponding to distinct eigenvalues are orthogonal. Note that when the Sturm-Liouville problem is regular, distinct eigenvalues are guaranteed. For background see [[Sturm-Liouville theory]].
Line 5: Line 7:
==Orthogonality Theorem==  
==Orthogonality Theorem==  


<math> \langle f, g\rangle = \int_{a}^{b} \overline{f(x)} g(x)w(x)\,dx</math> <math>=0</math>, where <math>f\left( x\right) </math> and <math>g\left( x\right) </math> are solutions to the Sturm-Liouville equation corresponding to distinct eigenvalues and  <math>w\left( x\right) </math> is the "weight" or "density" function.
<span style="display:inline-block; vertical-align:middle"><math> \langle f, g\rangle = \int_{a}^{b} \overline{f(x)} g(x)w(x)\,dx \ = \ 0</math> </span>, where f(x) and g(x) are solutions to the Sturm-Liouville equation corresponding to distinct eigenvalues and  w(x) is the "weight" or "density" function.


===Proof===  
===Proof===  


Let <math>f\left( x\right) </math> and  
Let f(x) and g(x) be solutions of the Sturm-Liouville equation [[Sturm-Liouville theory#(1) | (1) ]] corresponding to eigenvalues <math>\lambda </math> and <math> \mu </math> respectively. Multiply the equation for g(x) by  
<math>g\left( x\right) </math> be solutions of the Sturm-Liouville equation [[Sturm-Liouville theory#(1) | (1) ]] corresponding to eigenvalues <math>\lambda </math> and <math> \mu </math> respectively. Multiply the equation for <math>g\left( x\right) </math> by  
<span style="text-decoration:overline">f</span>(x) (the complex conjugate of f(x)) to get:
<math>\bar{f} \left( x\right) </math> (the complex conjugate of <math>f\left( x\right) </math>) to get:


<math>-\bar{f} \left( x\right) \frac{d\left( p\left( x\right) \frac{dg}{dx}
<span style="display:inline-block; vertical-align:middle"><math>-\bar{f} \left( x\right) \frac{d\left( p\left( x\right) \frac{dg}{dx}
\left( x\right) \right) }{dx} +\bar{f} \left( x\right) q\left( x\right)
\left( x\right) \right) }{dx} +\bar{f} \left( x\right) q\left( x\right)
g\left( x\right) =\mu  \bar{f} \left( x\right) w\left( x\right) g\left(
g\left( x\right) =\mu  \bar{f} \left( x\right) w\left( x\right) g\left(
x\right) </math> .
x\right) </math> </span>


(Only  
(Only  
<math>f\left( x\right) </math>, <math>g\left( x\right) </math>,  
f(x), g(x),  
<math>\lambda </math>, and  
<math>\lambda </math>, and  
<math>\mu </math>
<math>\mu </math>
may be complex; all other quantities are real.) Complex conjugate  
may be complex; all other quantities are real.) Complex conjugate  
this equation, exchange  
this equation, exchange  
<math>f\left( x\right) </math>
f(x)
and  
and  
<math>g\left( x\right) </math>, and subtract the new equation from the original:
g(x), and subtract the new equation from the original:




<math>-\bar{f} \left( x\right) \frac{d\left( p\left( x\right) \frac{dg}{dx}
<span style="display:inline-block; vertical-align:middle"><math>-\bar{f} \left( x\right) \frac{d\left( p\left( x\right) \frac{dg}{dx}
\left( x\right) \right) }{dx} +g\left( x\right) \frac{d\left( p\left(
\left( x\right) \right) }{dx} +g\left( x\right) \frac{d\left( p\left(
x\right) \frac{d\bar{f} }{dx} \left( x\right) \right) }{dx} =\frac{d\left(
x\right) \frac{d\bar{f} }{dx} \left( x\right) \right) }{dx} =\frac{d\left(
Line 35: Line 36:
x\right) -\bar{f} \left( x\right) \frac{dg}{dx} \left( x\right) \right]
x\right) -\bar{f} \left( x\right) \frac{dg}{dx} \left( x\right) \right]
\right) }{dx} =\left( \mu -\bar{\lambda} \right) \bar{f} \left( x\right)
\right) }{dx} =\left( \mu -\bar{\lambda} \right) \bar{f} \left( x\right)
g\left( x\right) w\left( x\right). </math>
g\left( x\right) w\left( x\right) </math></span> <br><br>
Integrate this between the limits  
Integrate this between the limits  
<math>x=a</math>
<math>x=a</math>
Line 54: Line 55:
conditions, which are either:
conditions, which are either:


: <math>\bullet </math> periodic boundary conditions, i.e., that <math>f\left( x\right) </math>, <math>g\left( x\right) </math>, and their first derivatives (as well as <math>p\left( x\right) </math>) have the same values at <math>x=b</math> as at <math>x=a</math>, or
: <math>\bullet </math> periodic boundary conditions, i.e., that f(x), g(x), and their first derivatives (as well as p(x)) have the same values at <math>x=b</math> as at <math>x=a</math>, or


: <math>\bullet </math> that independently at <math>x=a</math> and at <math>x=b</math> either:
: <math>\bullet </math> that independently at <math>x=a</math> and at <math>x=b</math> either:


:: <math>\bullet </math> the condition cited in equation [[Sturm-Liouville theory#(2) | (2) ]] or [[Sturm-Liouville theory#(3) | (3) ]] holds or:  
:: <math>\bullet </math> the condition cited in equation [[Sturm-Liouville theory#(2) | (2) ]] or [[Sturm-Liouville theory#(3) | (3) ]] holds or:  
:: <math>\bullet </math> <math>p\left( x\right) =0</math>.
:: <math>\bullet </math> <math>p(x)=0</math>.


So:  <math>\left( \mu -\bar{\lambda} \right) \int\nolimits_{a}^{b}\bar{f} \left(x\right) g\left( x\right) w\left( x\right) dx =0</math>.
So:  <span style="display:inline-block; vertical-align:middle"><math>\left( \mu -\bar{\lambda} \right) \int\nolimits_{a}^{b}\bar{f} \left(x\right) g\left( x\right) w\left( x\right) dx =0</math></span>


If we set  
If we set  
<math>f=g</math>
<math>f=g</math>
, so that the integral surely is non-zero, then it follows that  
, so that the integral surely is non-zero, then it follows that  
<math>\bar{\lambda} =\lambda </math>; that is, the eigenvalues are real, making the differential operator in the Sturm-Liouville equation self-adjoint (hermitian); so:
<span style="text-decoration:overline">&#955;</span> =&#955; that is, the eigenvalues are real, making the differential operator in the Sturm-Liouville equation self-adjoint (hermitian); so:


<math>\left( \mu -\lambda \right) \int\nolimits_{a}^{b}\bar{f} \left( x\right)
<span style="display:inline-block; vertical-align:middle"><math>\left( \mu -\lambda \right) \int\nolimits_{a}^{b}\bar{f} \left( x\right)
g\left( x\right) w\left( x\right) dx =0</math>
g\left( x\right) w\left( x\right) dx =0</math></span>
.


It follows that, if  
It follows that, if  
Line 77: Line 77:
<math>g</math>
<math>g</math>
have distinct eigenvalues, then they are orthogonal. QED.
have distinct eigenvalues, then they are orthogonal. QED.
===See also===
* [[Sturm-Liouville theory]]
* [[Normal mode]]
* [[Self-adjoint]]
===References===
1. Ruel V. Churchill, "Fourier Series and Boundary Value Problems", pp. 70-72, (1963) McGraw-Hill, ISBN 0-07-010841-2.
<!--
[[Category:Differential equations]]
[[Category:Ordinary differential equations]]
[[Category:Operator theory]]
[[Category:Linear operators]]
-->

Latest revision as of 19:34, 4 September 2009

This article is basically copied from an external source and has not been approved.
Main Article
Discussion
Related Articles  [?]
Bibliography  [?]
External Links  [?]
Citable Version  [?]
Proofs [?]
 
More information relevant to Sturm-Liouville theory.

This article proves that solutions to the Sturm-Liouville equation corresponding to distinct eigenvalues are orthogonal. Note that when the Sturm-Liouville problem is regular, distinct eigenvalues are guaranteed. For background see Sturm-Liouville theory.

Orthogonality Theorem

, where f(x) and g(x) are solutions to the Sturm-Liouville equation corresponding to distinct eigenvalues and w(x) is the "weight" or "density" function.

Proof

Let f(x) and g(x) be solutions of the Sturm-Liouville equation (1) corresponding to eigenvalues and respectively. Multiply the equation for g(x) by f(x) (the complex conjugate of f(x)) to get:

(Only f(x), g(x), , and may be complex; all other quantities are real.) Complex conjugate this equation, exchange f(x) and g(x), and subtract the new equation from the original:




Integrate this between the limits and


.

The right side of this equation vanishes because of the boundary conditions, which are either:

periodic boundary conditions, i.e., that f(x), g(x), and their first derivatives (as well as p(x)) have the same values at as at , or
that independently at and at either:
the condition cited in equation (2) or (3) holds or:
.

So:

If we set , so that the integral surely is non-zero, then it follows that λ =λ that is, the eigenvalues are real, making the differential operator in the Sturm-Liouville equation self-adjoint (hermitian); so:

It follows that, if and have distinct eigenvalues, then they are orthogonal. QED.