A functional model for commuting pairs of contractions and the symmetrized bidisc

Similar documents
A functional model for commuting pairs of contractions and the symmetrized bidisc

A functional model for commuting pairs of contractions and the symmetrized bidisc

Realization formulae for bounded holomorphic functions on certain domains and an application to the Carathéodory extremal problem

Characterizations of Some Domains via Carathéodory Extremals

Ando dilation and its applications

A generalized Schwarz lemma for two domains related to µ-synthesis

Complex geometry and operator theory

OPERATOR THEORY ON SYMMETRIZED BIDISC

ANDO DILATIONS, VON NEUMANN INEQUALITY, AND DISTINGUISHED VARIETIES

arxiv: v4 [math.cv] 7 Sep 2009

arxiv: v1 [math.cv] 9 Jan 2012

DILATIONS OF Γ-CONTRACTIONS BY SOLVING OPERATOR EQUATIONS

Rational and H dilation

ON THE UNILATERAL SHIFT AND COMMUTING CONTRACTIONS

ON THE SIMILARITY OF CENTERED OPERATORS TO CONTRACTIONS. Srdjan Petrović

Symmetric functions of two noncommuting variables

The spectral Carathéodory-Fejér problem

Schur functions. J. Rovnyak and H. S. V. de Snoo

Compression, Matrix Range and Completely Positive Map

arxiv:math/ v1 [math.cv] 27 Aug 2006

Generalized Numerical Radius Inequalities for Operator Matrices

A 3 3 DILATION COUNTEREXAMPLE

arxiv:math/ v1 [math.oa] 9 May 2005

Classical stuff - title to be changed later

SHARPENED FORMS OF A VON NEUMANN

Characterization of invariant subspaces in the polydisc

arxiv: v1 [math.cv] 7 Mar 2019

THE CENTRAL INTERTWINING LIFTING AND STRICT CONTRACTIONS

FUNCTION THEORY ON THE NEIL PARABOLA

Perturbations of Strongly Continuous Operator Semigroups, and Matrix Muckenhoupt Weights

Means of unitaries, conjugations, and the Friedrichs operator

Functional Analysis. Franck Sueur Metric spaces Definitions Completeness Compactness Separability...

Relaxed commutant lifting and a relaxed Nehari problem: Redheffer state space formulas

08a. Operators on Hilbert spaces. 1. Boundedness, continuity, operator norms

A von Neumann Wold decomposition of two-isometries

The Ultimate Estimate of the Upper Norm Bound for the Summation of Operators

REPRESENTATION THEORY WEEK 7

Factorizations of Kernels and Reproducing Kernel Hilbert Spaces

arxiv: v2 [math.fa] 17 May 2016

Weighted Commutant Lifting.

Elementary linear algebra

CONSERVATIVE DILATIONS OF DISSIPATIVE N-D SYSTEMS: THE COMMUTATIVE AND NON-COMMUTATIVE SETTINGS. Dmitry S. Kalyuzhnyĭ-Verbovetzkiĭ

Invariant subspaces for operators whose spectra are Carathéodory regions

( f(rz) 2 E. E E (D n ) := sup f(z) L(E,E ) <.

PAIRS OF COMMUTING ISOMETRIES - I

Operator positivity and analytic models of commuting tuples of operators

INSTITUTE OF MATHEMATICS THE CZECH ACADEMY OF SCIENCES. On ρ-dilations of commuting operators. Vladimír Müller

Lecture 5. Ch. 5, Norms for vectors and matrices. Norms for vectors and matrices Why?

HOLOMORPHIC MAPPINGS INTO SOME DOMAIN IN A COMPLEX NORMED SPACE. Tatsuhiro Honda. 1. Introduction

ON THE INDEX OF INVARIANT SUBSPACES IN SPACES OF ANALYTIC FUNCTIONS OF SEVERAL COMPLEX VARIABLES

Redheffer Representations and Relaxed Commutant Lifting

Asymptotic behaviour of Hilbert space operators with applications. Theses of Ph. D. dissertation. Supervisor:

arxiv: v1 [math.nt] 15 Sep 2009

ABSTRACT SCATTERING SYSTEMS: SOME SURPRISING CONNECTIONS

Hankel vector moment sequences

DILATION OF CONTRACTIVE TUPLES: A SURVEY

This is a repository copy of Geodesics, retracts, and the norm-preserving extension property in the symmetrized bidisc.

arxiv: v1 [math.fa] 24 May 2018

Adaptative Decomposition: The Case of the Drury Arveson Space

Lecture 5. Theorems of Alternatives and Self-Dual Embedding

Complex analysis techniques in the spectral theory of linear operators

REPRODUCING KERNEL FOR A CLASS OF WEIGHTED BERGMAN SPACES ON THE SYMMETRIZED POLYDISC

Bi-Isometries and Commutant Lifting

DILATIONS, WANDERING SUBSPACES, AND INNER FUNCTIONS

Asymptotic behaviour of Hilbert space contractions

COMMUTANT LIFTING FOR COMMUTING ROW CONTRACTIONS

Rotations & reflections

Quantum Computing Lecture 2. Review of Linear Algebra

INNER PRODUCT SPACE. Definition 1

The Knaster problem and the geometry of high-dimensional cubes

Review of Linear Algebra Definitions, Change of Basis, Trace, Spectral Theorem

INSTITUTE of MATHEMATICS. ACADEMY of SCIENCES of the CZECH REPUBLIC. Commutative dilation theory. Calin Ambrozie Vladimír Müller. Preprint No.

Topics in Operator Theory

Wold decomposition for operators close to isometries

Applied Analysis (APPM 5440): Final exam 1:30pm 4:00pm, Dec. 14, Closed books.

PICARD S THEOREM STEFAN FRIEDL

THE BOUNDEDNESS BELOW OF 2 2 UPPER TRIANGULAR OPERATOR MATRICES. In Sung Hwang and Woo Young Lee 1

Lecture # 3 Orthogonal Matrices and Matrix Norms. We repeat the definition an orthogonal set and orthornormal set.

CHAPTER 6. Representations of compact groups

On the Generalized Reid Inequality and the Numerical Radii

HARNACK AND SHMUL YAN PRE-ORDER RELATIONS FOR HILBERT SPACE CONTRACTIONS

MIT Algebraic techniques and semidefinite optimization May 9, Lecture 21. Lecturer: Pablo A. Parrilo Scribe:???

Commutant Lifting for Commuting Row Contractions

Introduction to finite element exterior calculus


Complex symmetric operators

Wavelets and Linear Algebra

On John type ellipsoids

Hilbert spaces. Let H be a complex vector space. Scalar product, on H : For all x, y, z H and α, β C α x + β y, z = α x, z + β y, z x, y = y, x

Projection Theorem 1

EXISTENCE OF NON-SUBNORMAL POLYNOMIALLY HYPONORMAL OPERATORS

MATH 423 Linear Algebra II Lecture 33: Diagonalization of normal operators.

On some interpolation problems

Zero weight spaces of irreducible representations and a new estimate of the isovariant Borsuk-Ulam constant for SU(3) Ikumitsu NAGASAKI 1)

SHARP BOUNDARY TRACE INEQUALITIES. 1. Introduction

Math Introduction to Numerical Analysis - Class Notes. Fernando Guevara Vasquez. Version Date: January 17, 2012.

= P HV n. T n 1 T m 2

The Choquet boundary of an operator system

Algebra I Fall 2007

Rotations & reflections

Transcription:

A functional model for commuting pairs of contractions and the symmetrized bidisc Nicholas Young Leeds and Newcastle Universities Lecture 3 A functional model for commuting pairs of contractions St Petersburg, June 2016

Γ-contractions recapitulation A commuting pair (S, P ) of operators on a Hilbert space H is a Γ-contraction if, for every polynomial g in two variables, g(s, P ) sup g. Γ If A, B are commuting contractions then (A + B, AB) is a Γ-contraction. If (S, P ) is a commuting pair of operators then (S, P ) is a Γ-contraction if and only if ρ(αs, α 2 P ) 0 for all α D, where ρ(s, P ) = 2(1 P P ) S + S P S + P S. By continuity for any Γ-contraction (S, P ) and any ω T, 2(1 P P ) ω(s S P ) ω(s P S) 0.

The fundamental operator of a Γ-contraction Let (S, P ) be a Γ-contraction on H. conjugate by DP 1 obtain If P < 1, we may = (1 P P ) 1 2 in the last inequality to 1 Re{ωDP 1 (S S P )DP 1 } 0 for all ω T. Hence the operator F = DP 1 (S S P )DP 1 B(D P ) satisfies w(f ) 1. For a general Γ-contraction (S, P ) there exists a unique F B(D P ) such that S S P = D P F D P, and this unique F satisfies w(f ) 1. This F is called the fundamental operator of the Γ-contraction (S, P ) [4].

A functional model for pure Γ-contractions Theorem (Bhattacharyya and Pal, 2014) Let (S, P ) be a pure Γ-contraction (so that P is a pure contraction). Let F be the fundamental operator of (S, P ). Then (S, P ) is unitarily equivalent to the pair (S, P) on the Hilbert space H P def = H 2 (D P ) Θ P H 2 (D P ) where (S, P) is the compression of the Γ-isometry ( T F +zf, T z ) on H 2 (D P ) to its co-invariant subspace H P. Observe that P is precisely the Nagy-Foias model of P.

Proof Recall the proof of the Nagy-Foias model. We use Θ P (λ) = [ P + λd P (1 λp ) 1 D P ] D P. The map given by U : H H P def = H 2 (D P ) Θ P H 2 (D P ) is a unitary operator, and (Ux)(z) = D P (1 H zp ) 1 x UP U = the compression to H P of T z 1 DP = P.

Proof 2 Since F is the fundamental operator of (S, P ), we have F B(D P ), w(f ) 1, and S SP = D P F D P. Since w(f ) 1, the pair ( T F +zf, T z ) is a Γ-isometry on H 2 (D P ). We shall next prove that W T F +zf = SW. Observe that W : H 2 (D P ) H satisfies W (z n y), x = z n y, DP x + zd P P x +... = y, D P P n x = P n D P y, x. Thus W (z n y) = P n D P y.

Proof 3 Hence W T F +zf (z n y) = W (z n F y + z n+1 F y) = P n D P F y + P n+1 D P F y, while SW z n y = SP n D P y = P n SD P y. Therefore (W T F +zf SW )(z n y) = P n Hy, where H = D P F + P D P F SD P : DP H.

Proof 4 H = D P F + P D P F SD P : D P H. Now HD P = D P F D P + P D P F D P SDP 2 = (S P S ) + P (S SP ) S(1 P P ) = 0. Hence H = 0, and so W T F +zf = SW. Accordingly (T F +zf ) W = W S.

Proof 5 The equation (T F +zf ) W = W S implies that (T F +zf ) H P = T F +zf ran W ran W = H P. Thus H P is a co-invariant subspace of H 2 (D P ) for T F +zf. In the equation (T F +zf ) W = W S : H H 2 (D P ) take the range restriction U of W to obtain (T F +zf ) U = US : H H P.

Proof 6 Hence (T F +zf ) H P = US U : H P H P. Finally USU = the compression of T F +zf to H P = S. We have shown that U(S, P )U = (S, P).

Completely non-unitary Γ-contractions Every Γ-contraction is the orthogonal direct sum of a Γ- unitary and a completely non-unitary Γ-contraction, meaning one for which the second component is a completely non-unitary contraction. Let (S, P ) be a Γ-contraction on H. Let H 1 be the maximal subspace of H on which P is unitary and let H 2 = H H 1. Then (1) H 1, H 2 reduce both S and P (2) (S H 1, P H 1 ) is Γ-unitary (3) (S H 2, P H 2 ) is a Γ-contraction and P H 2 is completely non-unitary.

A functional model for Γ-contractions Theorem (Sarkar 2015) Let (S, P ) be a completely non-unitary Γ-contraction. Let H P be the Nagy-Foias model space of the contraction P : H P = [ H 2 (D P ) P L 2 (D P ) ] { Θ P u P u : u H 2 (D P ) }. Then (S, P ) is unitarily equivalent to (S, P), where P is the Nagy-Foias model of P and S is the compression to H P of the operator T F +zf U on H 2 (D P ) P L 2 (D P ), where F is the fundamental operator of (S, P ) and U B( P L 2 (D P )) is such that (U, M e it) is Γ-unitary on P L 2 (D P ).

References [1] B. Sz.-Nagy and C. Foias, Harmonic Analysis of operators on Hilbert space, Akadémiai Kiadő, Budapest 1970. [2] J. Agler and N. J. Young, A model theory for Γ-contractions, J. Operator Theory 49 (2003) 45-60. [3] T. Bhattacharyya and S. Pal, A functional model for pure Γ-contractions, J. Operator Theory 71 (2014). [4] T. Bhattacharyya, S. Pal and S. Shyam Roy, Dilation of Γ-contractions by solving operator equations, Advances in Mathematics 230 (2012) 577-606. [5] J. Sarkar, Operator theory on symmetrized bidisc, Indiana Univ. Math. J. 64 (2015) 847 873.

More about the symmetrized bidisc The domain G def = {(z + w, zw) : z < 1, w < 1}, the interior of Γ, is called the open symmetrized bidisc. There are at least four reasons to take an interest in G, connected to classical complex function theory of the unit disc D the Nagy-Foias theory of models of operators the theory of invariant distances in several complex variables a problem in robust control theory.

The 2 2 spectral Nevanlinna-Pick problem Given distinct points λ 1,..., λ N D and square matrices W 1,..., W N C 2 2, none of them a scalar multiple of the identity, ascertain whether there exists an analytic function F : D C 2 2 such that F (λ j ) = W j for j = 1,..., N and r(f (λ)) 1 for all λ D where r( ) is the spectral radius. This problem is equivalent to an interpolation problem for analytic functions f : D G. This is because r(w ) 1 if and only if (tr W, det W ) Γ.

A Schwarz-Pick Lemma for the spectral radius Let Ω be the set of 2 2 matrices of spectral radius < 1. Let W 1, W 2 Ω be non-scalar matrices and let λ 1, λ 2 D. Let s j = tr W j, p j = det W j for j = 1, 2. There exists an analytic function F F (λ j ) = W j for j = 1, 2 if and only if sup ω =1 (s 2 p 1 s 1 p 2 )ω 2 + 2(p 2 p 1 )ω + s 1 s 2 (s 1 s 2 p 1 )ω 2 2(1 p 2 p 1 )ω + s 2 p 2 s 1 : D Ω such that λ 1 λ 2 1 λ 2 λ 1. This result comes out of the study of the Carathéodory and Kobayashi distances on G.

The Γ-Nevanlinna-Pick problem Given distinct points λ 1,..., λ N D and points (s 1, p 1 ),..., (s N, p N ) in Γ, determine whether there exists an analytic function h : D Γ such that h(λ j ) = (s j, p j ) for j = 1,..., N. Is there a Pick-type criterion for solvability? Answer: almost certainly not. We can solve the Γ-Nevanlinna-Pick problem in the case that N = 2.

Weak solvability Say that a Γ-Nevanlinna-Pick problem λ j D (s j, p j ) G for j = 1,..., N (1) is weakly solvable if, for every analytic map g : G D, the (classical) Nevanlinna-Pick problem λ j g(s j, p j ) is solvable. Clearly solvable problems are weakly solvable (if h satisfies conditions (1) then g h maps λ j to g(s j, p j )). Theorem. The Γ-Nevanlinna-Pick problem (1) is weakly solvable if and only if, for all ω T, [ ] (2 ω si )(2 ωs j ) (2 ω p i s i )(2ωp j s j ) N 1 λ i λ j i,j=1 0.

Two-point Γ-Nevanlinna-Pick problems Theorem. Let λ 1, λ 2 be distinct points in D and let (s 1, p 1 ), (s 2, p 2 ) belong to G. The Γ-Nevanlinna-Pick problem λ j D (s j, p j ) for j = 1, 2 (2) is solvable if and only if it is weakly solvable. Corollary. The Γ-Nevanlinna-Pick problem (2) is solvable if and only if, for all ω T, [ ] (2 ω si )(2 ωs j ) (2 ω p i s i )(2ωp j s j ) 2 1 λ i λ j i,j=1 0. For the three-point Γ-Nevanlinna-Pick problem, weak solvability does not imply solvability.

The Carathéodory distance Let d denote the hyperbolic distance on D: for λ, µ D, d(λ, µ) def = tanh 1 λ µ 1 µλ. For any bounded domain Ω C n, the Carathéodory distance on Ω is defined for any points z, w Ω by C Ω (z, w) def = sup{d(g(z), g(w)) : g is analytic from Ω to D}. Any analytic function g : Ω D for which the supremum is attained is called a Carathéodory extremal function for z, w in Ω.

The Carathéodory distance on G The above solvability criterion for two-point Γ-Nevanlinna- Pick problems translates into the following statement. Theorem For any pair of points z 1 = (s 1, p 1 ), z 2 = (s 2, p 2 ) in G, there is a Carathéodory extremal function of the form Φ ω for some ω T, and therefore C G (z 1, z 2 ) = sup d(φ ω (z 1 ), Φ ω (z 2 )) ω =1 = tanh 1 sup ω =1 (s 2 p 1 s 1 p 2 )ω 2 + 2(p 2 p 1 )ω + s 1 s 2 (s 1 s 2 p 1 )ω 2 2(1 p 2 p 1 )ω + s 2 p 2 s 1.

The Kobayashi distance For any bounded domain Ω C n, the Lempert function on Ω is defined for any points z, w Ω to be δ Ω (z, w) def = inf d(λ, µ) over all λ, µ D such that there exists an analytic map f : D Ω such that f(λ) = z and f(µ) = w. δ Ω does not in general satisfy the triangle inequality. The Kobayashi distance K Ω on Ω is defined to be the greatest distance K on Ω such that K δ Ω. Since C Ω δ Ω, always C Ω K Ω.

The Kobayashi distance on G Lempert s theorem states that, for any bounded convex domain Ω, C Ω = K Ω = δ Ω. C. Costara proved that G is not isomorphic to any convex domain. Nevertheless: Theorem. C G = K G = δ G. This theorem solved a long-standing problem in the theory of invariant distances: are all domains for which C = K isomorphic to convex domains?

References J. Agler and N. J. Young, The hyperbolic geometry of the symmetrised bidisc, J. Geometric Analysis 14 (2004) 375-403. N. J. Young, Some analysable instances of mu-synthesis, Mathematical methods in systems, optimization and control, Editors: H. Dym, M. de Oliveira, M. Putinar, Operator Theory: Advances and Applications 222 349 366, Birkhäuser, Basel, 2012.