Bounded Borel functional calculus

from Wikipedia, the free encyclopedia

The bounded Borel functional calculus is a tool for studying Von Neumann algebras .

This functional calculus is an extension of the continuous functional calculus known from the theory of C * -algebras to limited Borel functions . This extension of the functional calculus is not possible in general C * -algebras, one has to restrict oneself to the smaller class of Von Neumann algebras.

construction

If one considers a bounded, monotonically growing sequence of continuous real-valued functions that are defined on the spectrum of a normal element of a Von Neumann algebra ( Hilbert space ), the pointwise limit is generally not continuous again. In is the sequence , where the continuous functional calculus is formed, a bounded and monotonically increasing (for the arrangement see positive operator ) sequence of self-adjoint operators , of which one can show that it converges in the strong operator topology . Since Von-Neumann algebras are exactly the sub-C * algebras of with one element that are closed in the strong operator topology, this limit value is again in .

If there is another sequence of continuous real-valued functions that converges monotonically to point by point , then one can show that the limit values ​​of and agree. Therefore it makes sense to designate this limit value with .

If the limit function is even continuous, then according to Dini's theorem , there is uniform convergence, and one recognizes that the definition just made is compatible with the continuous functional calculus. A continuation of these ideas leads to the so-called restricted Borel functional calculus (or Borel calculus for short).

The limited Borel calculus

If a normal element of a Von Neumann algebra and denotes the algebra of the Borel functions defined on , then the following applies:

  • There is exactly one * -homomorphism with , and the following continuity property: If the sequence of real-valued functions converges monotonously to in point by point , then the supremum of is in the Von Neumann algebra .

The suggestive spelling is used . The following can be shown:

  • There are formulas , for all .
  • Applies to everyone .
  • Is and so is true .
  • for everyone .
  • The restriction to the algebra of continuous functions is the continuous functional calculus .

A spectral mapping theorem cannot apply since the image of the spectrum under a Borel function is generally not compact again.

This functional calculus of bounded Borel functions is closely related to the spectral theorem . If, for example, self-adjoint, then the associated spectral family is, where the characteristic function denotes.

Applications

As an application, it should only be mentioned that this functional calculus leads to the construction of very many projections in Von Neumann algebras. If is a Borel set and denotes the associated characteristic function , then applies . Hence , that is, is an orthogonal projection in .

Since continuous functions can be approximated uniformly by simple functions , one can see with the help of the functional calculus that every element of a Von Neumann algebra is a norm limit of linear combinations of orthogonal projections . In this sense, there are many projections in Von Neumann algebras. In this way, the Von Neumann theory differs considerably from the theory of C * algebras. The C * algebra of continuous functions on the interval [0,1] has the 0 and 1 function as the only projections, and thus as few projections as possible.

This richness of projections is one of the main starting points of the theory of von Neumann algebras, so are factors , for example, by the structure of their projection associations classified .

literature