Compact Operators
5.1 Definition
A linear operator is compact if the image of the closed unit ball, , is relatively compact (its closure is compact) in .
Proposition 5.1. Every compact operator is bounded. Every finite-rank operator is compact.
Proposition 5.2. If is compact and is bounded, then and are compact.
Proposition 5.3. If are compact and in operator norm, then is compact.
5.2 Spectral Theory for Compact Operators
Theorem 5.4 (Spectral Theorem for Compact Self-Adjoint Operators). Let be a compact self-adjoint operator on a Hilbert space . Then:
- All eigenvalues of are real.
- Eigenvectors corresponding to distinct eigenvalues are orthogonal.
- There exists an orthonormal basis of consisting of eigenvectors of .
- If are the nonzero eigenvalues with orthonormal eigenvectors , then .
Corollary 5.5. A compact self-adjoint operator on an infinite-dimensional Hilbert space has at most countably many nonzero eigenvalues, and is the only possible accumulation point.
5.3 Spectral Theorem for Normal Operators
A bounded operator on a Hilbert space is normal if , and unitary if .
Proposition 5.6. If is normal, then for all .
Proposition 5.7. If is normal, then and eigenvectors corresponding to distinct eigenvalues are orthogonal.
Theorem 5.8 (Spectral Theorem for Normal Compact Operators). Let be a compact normal operator on a Hilbert space . Then there exists an orthonormal basis of consisting of eigenvectors of , and the eigenvalues satisfy .
This generalises Theorem 5.4: self-adjoint operators are normal, and unitary operators are normal (with eigenvalues on the unit circle in ).
Theorem 5.9 (Spectral Theorem for Bounded Normal Operators). Let be a bounded normal operator on . There exists a unique projection-valued measure on the Borel subsets of such that
This integral representation implies: if is a bounded Borel function on , then defines a bounded normal operator satisfying the functional calculus relations.
5.4 Fredholm Alternative
Theorem 5.6 (Fredholm Alternative). Let be a compact operator on a Banach space and . Then exactly one of the following holds:
- is bijective (hence invertible by the bounded inverse theorem).
- has a nontrivial solution (i.e., is an eigenvalue of ).