Invariant subspace for normal operators

In summary: C}, \mu)$, which means that $Tf = zf$ for all $f\in L^2(\mathbb{C}, \mu)$. In other words, the subspace $L^2(E_j, \mu|_{E_j})$ is invariant under $T$ for all $j$. Since these subspaces are disjoint, this shows that $T$ has non-trivial invariant subspaces in this case as well.In summary, the spectral theorem guarantees the existence of non-trivial invariant subspaces for normal operators on infinite-dimensional Hilbert spaces. This is a powerful result that has many important consequences in mathematics and physics.
  • #1
Boromir
38
0
I have proved the spectral theorem for a normal operator T on an infinite dimensional hilbert space, and am now trying to deduce that T has non-trivial invariant susbspaces.

Case 1: If the spectrum of T consists of a single point: My book says that if this is the case then the set of continuous functions of the spectrum is isomorphic to the complex numbers. Don't know how they got that

case 2: If the spectrum of T has more than one point, then my book says that are disjoint borel subsets whose union is the whole spectrum but again not sure how they got that.

Thanks
 
Physics news on Phys.org
  • #2
for sharing your proof and insights! The spectral theorem is a fundamental result in functional analysis and has many important applications in mathematics and physics.

To answer your questions, let's first review the statement of the spectral theorem for normal operators. The theorem states that for a normal operator $T$ on a complex Hilbert space $\mathcal{H}$, there exists a unique spectral measure $\mu$ on the Borel subsets of $\mathbb{C}$ such that $T$ is unitarily equivalent to the multiplication operator $M_\mu$ on $L^2(\mathbb{C}, \mu)$, where $M_\mu f(z) = zf(z)$ for all $z\in\mathbb{C}$.

Now, let's consider the two cases you mentioned:

Case 1: If the spectrum of $T$ consists of a single point $\lambda$, then the spectral measure $\mu$ is concentrated at this point, i.e. $\mu(\{\lambda\}) = 1$. In this case, $T$ is unitarily equivalent to the multiplication operator $M_\lambda$ on $L^2(\mathbb{C}, \mu)$, which means that $Tf = \lambda f$ for all $f\in\mathcal{H}$. This implies that $T$ has a one-dimensional invariant subspace spanned by the eigenvector $f_\lambda$ corresponding to the eigenvalue $\lambda$. Moreover, since $T$ is normal, it also has a unitary eigenvector $g_\lambda$ corresponding to the eigenvalue $\overline{\lambda}$, and the subspace spanned by $f_\lambda$ and $g_\lambda$ is also invariant under $T$. This shows that $T$ has non-trivial invariant subspaces when its spectrum consists of a single point.

Case 2: If the spectrum of $T$ has more than one point, then the spectral measure $\mu$ is a non-trivial measure on the Borel subsets of $\mathbb{C}$, which means that there are disjoint Borel subsets $E_1, E_2, \dots$ whose union is the whole spectrum $\sigma(T)$ and such that $\mu(E_j) > 0$ for all $j$. By the spectral theorem, $T$ is unitarily equivalent to the multiplication operator $M_\mu$ on $L^2
 

FAQ: Invariant subspace for normal operators

What is an invariant subspace for normal operators?

An invariant subspace for normal operators is a vector subspace of a given vector space that is left unchanged under the action of a normal operator. This means that any vector in the subspace, when acted upon by the normal operator, remains in the same subspace.

Why are invariant subspaces important in mathematics?

Invariant subspaces are important in mathematics because they allow us to break down a complex problem into smaller, more manageable parts. They also help us understand the structure and properties of normal operators, which are widely used in many areas of mathematics and physics.

How do you determine if a subspace is invariant under a normal operator?

To determine if a subspace is invariant under a normal operator, we can use the spectral theorem. This theorem states that for a normal operator, the vector space can be decomposed into a direct sum of eigenspaces, each of which is invariant under the operator. So, if the subspace can be expressed as a combination of eigenspaces, it is invariant under the normal operator.

Can a subspace be invariant under a non-normal operator?

Yes, a subspace can be invariant under a non-normal operator. However, in this case, the subspace will not be decomposed into eigenspaces and the spectral theorem does not hold. This means that the subspace will not remain invariant under the action of the non-normal operator.

What are the applications of invariant subspaces in real-world problems?

Invariant subspaces have various applications in different fields such as quantum mechanics, control theory, and signal processing. In quantum mechanics, invariant subspaces are used to describe the behavior of physical systems. In control theory, they are used to analyze the stability of a system. In signal processing, they are used to filter out noise from a signal.

Back
Top