$$\mu$$ -Norm of an Operator

  • PDF / 367,230 Bytes
  • 29 Pages / 612 x 792 pts (letter) Page_size
  • 46 Downloads / 193 Views

DOWNLOAD

REPORT


Norm of an Operator D. V. Treschev a Received January 17, 2020; revised January 17, 2020; accepted April 8, 2020

Abstract—Let (X , μ) be a measure space. For any measurable set Y ⊂ X let 1Y : X → R be the indicator of Y and let πY be the orthogonal projection L2 (X )  f → πY f = 1Y f . For  any bounded operator W on L2 (X , μ) we define its μ-norm W μ = inf χ μ(Yj )W πY 2 , where the infimum is taken over all measurable partitions χ = {Y1 , . . . , YJ } of X . We present some properties of the μ-norm and some computations. Our main motivation is the problem of constructing a quantum entropy. DOI: 10.1134/S008154382005020X

1. INTRODUCTION AND MOTIVATION Let X be a nonempty set and let B be a σ-algebra of subsets X ⊂ X . Consider the measure space (X , B, μ), where μ is a probability measure: μ(X ) = 1. Consider the Hilbert space H = L2 (X , μ) with the scalar product and norm   f  = f, f . f, g = f g dμ, X

For any bounded operator W on H, let W  be its norm defined by W  = sup W f . f =1

For any X ∈ B consider the orthogonal projection πX : H → H,

H  f → πX f = 1X · f,

(1.1)

where 1X is the indicator of X (the function equal to 1 on X and vanishing at other points). Then μ(X) > 0 implies πX  = 1, and for any X  , X  ∈ B πX  + πX  = πX  ∪X  + πX  ∩X  ,

πX  πX  = πX  ∩X  .

We say that χ = {Y1 , . . . , YJ } is a (finite measurable) partition of X if    μ X\ Yj = 0, μ(Yj ∩ Yk ) = 0 for any j, k ∈ {1, . . . , J}, k = j. Yj ∈ B, 1≤j≤J

For any two partitions χ = {Y1 , . . . , YJ } and κ = {X1 , . . . , XK }, we define χ ∨ κ = {Yj ∩ Xk }j=1,...,J, k=1,...,K . Obviously, χ ∨ κ is also a partition. a Steklov Mathematical Institute of Russian Academy of Sciences, ul. Gubkina 8, Moscow, 119991 Russia.

E-mail address: [email protected]

262

μ-NORM OF AN OPERATOR

263

Let W be a bounded operator on H. For any partition χ = {Y1 , . . . , YJ } of X we define Mχ (W ) =

J 

μ(Yj )W πYj 2 .

(1.2)

j=1

We define the μ-norm1 of W by W μ = inf



χ

Mχ (W ).

(1.3)

Recall that the operator U is said to be an isometry if f, g = U f, U g,

f, g ∈ H.

If the isometry U is invertible, then U is called a unitary operator. For any bounded W , any Y ∈ B, and any isometry U , W πY  ≤ W ,

U W  = W .

This implies the following obvious properties of the μ-norm: W μ ≤ W ,

idμ = 1,

(1.4)

W1 W2 μ ≤ W1  · W2 μ , λW μ = |λ| · W μ W μ = U W μ

(1.5)

for any λ ∈ C,

(1.6)

for any unitary U.

(1.7)

The first question which probably comes to the reader’s mind is “why.” Why such a construction may be useful or interesting? Now we are going to explain our motivations. Let F : X → X be an endomorphism of the measure space (X , B, μ). This means that for any X ∈ B the set F −1 (X) (complete preimage) also lies in B and μ(X) = μ(F −1 (X)). Invertible endomorphisms are called automorphisms. Let End(X ) denote the semigroup of all endomorphisms of (X , B, μ). There are two standard constructions associated with any F ∈ End(X , μ): (1) any such F generates an i