Definition
Rigged Hilbert space
A dense nuclear test space inside a Hilbert space, together with its continuous anti-dual.
Definition
A rigged Hilbert space, or Gelfand triple, is a chain
in which is a Hilbert space, is a Hausdorff nuclear locally convex space continuously and densely embedded in , and is its continuous anti-dual with the strong dual topology. Taking the inner product on linear in its first variable, the second embedding sends to the conjugate-linear functional . Density makes this embedding injective.
Test vectors and generalized vectors
The space carries a topology finer than the Hilbert norm topology, encoding regularity or decay. Elements of act as generalized vectors, including distributions that do not belong to . The Hilbert space lies between these two scales and identifies ordinary vectors with continuous anti-linear functionals on the test space.
Operators and spectral analysis
If a densely defined operator on preserves and acts continuously on its locally convex topology, duality extends it to . Generalized eigenvectors may then live in even when the operator has no eigenvectors in . Nuclearity supplies the compactness and kernel-theorem structure used in generalized spectral expansions Gel′fand–Vilenkin, Chapter I.
Example and convention
The canonical example is
with Schwartz functions as test vectors and tempered distributions as generalized vectors. Some authors use rigged Hilbert space for any continuous dense embedding , without requiring to be nuclear, and reserve nuclear Gelfand triple for the definition above. The nuclear convention is used here.
References
- I. M. Gel′fand and N. Ya. Vilenkin, Generalized Functions, Volume 4: Applications of Harmonic Analysis, Academic Press, 1964; AMS Chelsea reprint. AMS publisher record. Relevant: Chapter I on nuclear spaces, rigged Hilbert spaces, and spectral analysis.
- Michael Reed and Barry Simon, Methods of Modern Mathematical Physics, Volume I: Functional Analysis, Academic Press, 1972. Elsevier DOI record. Relevant: Chapter V on locally convex spaces and distributions.