Figuring symmetries of a differential operator from its eigenfunctions

  • Thread starter JPaquim
  • Start date
  • #1
JPaquim
34
0
So, I understand that the derivative operator, [itex]D=\frac{d}{dx}[/itex] has translational invariance, that is: [itex]x \rightarrow x - x_0[/itex], and its eigenfunctions are [itex]e^{\lambda t}[/itex]. Analogously, the theta operator [itex]\theta=x\frac{d}{dx}[/itex] is invariant under scalings, that is [itex]x \rightarrow \alpha x[/itex], and its eigenfunctions are [itex]x^\lambda[/itex]. Taking logarithms and exponentials, I have constructed a sequence of operators and their respective eigenfunctions, all with the property that [itex]\{L(\frac{d}{dx})\}f^\lambda(x)=\lambda f^\lambda(x)[/itex]. I've taken a picture and attached it to this post.

My guess is that associated with every single one of these operators is some symmetry, some sort of coordinate transformation [itex]x \rightarrow f(x)[/itex] under which the operator is invariant. For the [itex]x\log x \frac{d}{dx}[/itex] operator, its invariant under [itex]x \rightarrow x^k[/itex], by inspection. How can I figure out what sort of symmetry a given operator has, given its eigenfunctions?

Physically, symmetries are associated with conservation laws. For a system whose differential equations are governed by this sort of differential operators, what sort of conserved quantities should I expect?
 

Attachments

  • IMG_0596.jpg
    IMG_0596.jpg
    16.5 KB · Views: 375

Answers and Replies

  • #2
The_Duck
1,006
108
##\frac{d}{dx}## is associated with translational symmetry. An infinitesimal translation is produced by acting on a function with the operator ##1 + \epsilon \frac{d}{dx}##, with ##\epsilon## infinitesimal.

##x \frac{d}{dx}## is associated with scale invariance. An infinitesimal rescaling is produced by acting on a function with the operator ##1 + \epsilon x \frac{d}{dx}##, with ##\epsilon## infinitesimal.

Presumably your other operators ##O## can be associated with symmetry transformations with an infinitesimal transformation being implemented by ##1 + \epsilon O##?
 
  • #3
JPaquim
34
0
Ok, I agree with you. How can I figure out the "finite" version of the transformation from its infinitesimal counterpart?
 

Suggested for: Figuring symmetries of a differential operator from its eigenfunctions

Replies
14
Views
498
Replies
2
Views
695
  • Last Post
Replies
1
Views
452
Replies
2
Views
336
  • Last Post
Replies
2
Views
296
Replies
1
Views
184
  • Last Post
Replies
2
Views
668
  • Last Post
Replies
8
Views
624
Top