Proving No Eigenvalues Exist for an Operator on a Continuous Function Space

member 428835
Homework Statement
Show no eigenvalues exist for the operator ##Af(x) = xf(x)## where ##A:C[0,1]\to C[0,1]##.
Relevant Equations
Nothing comes to mind.
Eigenvalues ##\lambda## for some operator ##A## satisfy ##A f(x) = \lambda f(x)##. Then

$$
Af(x) = \lambda f(x) \implies\\
xf(x) = \lambda f(x)\implies\\
(\lambda-x)f(x) = 0.$$

How do I then show that no eigenvalues exist? Seems obvious one doesn't exists since ##\lambda-x \neq 0## for all ##x\in [0,1]##.
 
Physics news on Phys.org
Well the argument is subtle.

Since ##x## ranges over every point in ##[0,1]## and the supposed eigenvalue is some constant you cannot have ##\lambda=x##, since the function ##x## is not constant.
And obviously ##f(x)\ne 0 ##, otherwise it wouldn't be an eigenvalues' problem.
 
You are in fact almost there. The condition
##(\lambda-x)f(x) =0## implies that we have ##f(x) = 0## for all ##x\neq \lambda##. By continuity however ##f=0##. Thus, no eigenvector exists, as any possible eigenvalue leads to the trivial solution.
 
Last edited by a moderator:
  • Like
Likes member 428835
Math_QED said:
You are in fact almost there. The condition
##(\lambda-x)f(x) =0## implies that we have ##f(x) = 0## for all ##x\neq \lambda##. By continuity however ##f=0##. Thus, no eigenvector exists, as any possible eigenvalue leads to the trivial solution.
This just seems so simple. See, the HW was assigned in three parts, where part b) asked to show 0 is not an eigenvalue and part c) asked to prove no eigenvalues exist. But if this is part c), then why bother doing b)? Have I missed anything about ##\lambda = 0##? It seems this logic holds for all ##\lambda##.
 
joshmccraney said:
This just seems so simple. See, the HW was assigned in three parts, where part b) asked to show 0 is not an eigenvalue and part c) asked to prove no eigenvalues exist. But if this is part c), then why bother doing b)? Have I missed anything about ##\lambda = 0##? It seems this logic holds for all ##\lambda##.

It also works for ##\lambda = 0##. Then your equation leads to ##xf(x) = 0 \forall x## and for ##x \neq 0## we then have ##f(x) = 0##. By continuity again also ##f(0) = 0##, so ##f=0## everywhere.

So, to answer your question concretely, I don't think doing part (b) separately is necessary.

Note that eigenvalue ##\lambda = 0## can only happen when ##A## is non-injective. This map is injective however (continuity is key again here). Maybe that's what (b) wants you to write?
 
  • Like
Likes member 428835
There are two things I don't understand about this problem. First, when finding the nth root of a number, there should in theory be n solutions. However, the formula produces n+1 roots. Here is how. The first root is simply ##\left(r\right)^{\left(\frac{1}{n}\right)}##. Then you multiply this first root by n additional expressions given by the formula, as you go through k=0,1,...n-1. So you end up with n+1 roots, which cannot be correct. Let me illustrate what I mean. For this...
Back
Top