Schur parametrizations of stable all-pass discrete-time systems and balanced realizations. Application to rational L 2 approximation.

Size: px
Start display at page:

Download "Schur parametrizations of stable all-pass discrete-time systems and balanced realizations. Application to rational L 2 approximation."

Transcription

1 Schur parametrizations of stable all-pass discrete-time systems and balanced realizations. Application to rational L 2 approximation. Martine Olivi INRIA, Sophia Antipolis 12 Mai

2 Outline. Introduction : lossless functions and rational L 2 approximation Schur parametrizations and balanced realizations The software RARL2 New atlases 2

3 Lossless functions: transfer functions of stable allpass systems. Lossless function (discrete-time): G(z) square rational stable such that with equality on the unit circle. G(z) G(z) I, z > 1 A lossless function G(z) has a minimal realization s.t. (D, C, B, A) balanced R = D C unitary. B A lossless transfer functions in discrete-time can be represented by unitary realization matrices. Möbius transform: continuous-time dicrete-time. 3

4 Lossless functions in linear systems theory. Douglas Shapiro Shields factorization: S rational stable S = P G G lossless, P analytic in the unit disk (unstable). Many problems in linear system theory can be stated in terms of lossless functions : stochastic identification (Separable Least Squares) model reduction problems via rational approximation control design? 4

5 Rational L 2 approximation. H 2 Hardy space of matrix valued functions analytic in D, H 2 Hardy space of matrix valued functions analytic outside D, vanishing at F 2 = 1 { 2π 2π Tr 0 } F (e it )F (e it ) dt. Given F H 2, minimize F H 2 H rational, stable of McMillan degree n. Applications developped in the APICS team: identification from partial frequency data (hyperfrequency filters) sources and cracks detection (MEG, EEG) 5

6 The criterion. H = P G a best approximant of F : P proj. of F G 1 onto H 2 For G lossless (up to a left unitary factor) J(G) = F P (G) G 2. F (z) = C(zI N A) 1 B + D, H(z) = γ(zi n A) 1 B + D. For (A, B) input-normal pair (AA + BB = I) J(A, B) = F 2 Tr (γγ ), where γ = CW, W solution to AW A + BB = W. 6

7 Parametrization issue in the MIMO case In view of optimization, it is better use an atlas of charts attached to a differential manifold M a collection of local coordinate maps (charts) (W i, ϕ i ) W i open in M, ϕ i : W i R d diffeomorphism their domains cover the manifold: W i = M the changes of coordinates ϕ ij are smooth 7

8 Advantages of such a parametrization... in the representation of stable allpass systems: ensures identifiability a small perturbation of the parameters preserves the stability and the order of the system allows for the use of differential tools: a search algorithm can be run over the manifold changing from one local coordinate map to another when necessary. 8

9 Two approaches. L p n: p p (complex) lossless functions of McMillan degree n is a real differentiable manifold of dimension 2np + p 2 Subclass of real functions G( z) = G(z) dim np + p(p 1)/2 charts from realizations theory (state-space) M. Hazewinkel, R.E. Kalman (1974) for proper linear systems using nice selections B. Hanzon, R.J. Ober (1998) charts from Schur type algorithms (functional) D. Alpay, L. Baratchart, A. Gombani (1994) interpolation theory, reproducing kernel Hilbert spaces Ball, Gohberg, Rodman, Dym, Alpay,... An atlas which combines these two approaches B. Hanzon, M.O., R. Peeters (1999) 9

10 The SISO case : a balanced canonical form. A lossless function of degree n has a (unique) realization such that (i) (A, b, c, d) balanced [b, Ab, A 2 b,..., A n 1 b] positive upper triangular or equivalently (ii) R = d c is unitary and positive upper-hessenberg. b A B. Hanzon, R. Peeters (2000) 10

11 R = Then d b c A Parametrization. is unitary and positive upper-hessenberg. γ n κ n 0 κ n γ n I n 1 d c A 0 }{{} R = R n 1, γ n < 1, κ n = 1 γ n 2, γ n = d. By induction: (R k ) k=n,...,0 unitary upper-hessenberg order k and (γ k ) k=n,...,0, R 0 = γ 0, γ 0 = 1. 11

12 Hessenberg form. γ n κ n γ n 1 κ n κ n 1 γ n 2... κ n κ n 1... κ 1 γ 0 κ n γ n γ n 1 γ n κ n 1 γ n 2... γ n κ n 1... κ 1 γ 0 0 κ n 1 γ n 1 γ n 2... γ n 1 κ n 2... κ 1 γ κ n 2 γ n 2 κ n 3... κ 1 γ κ 2 γ 2 γ 1 γ 2 κ 1 γ κ 1 γ 1 γ 0 (1) 12

13 Connection with the Schur algorithm. Let g k (z) = γ k + c k (zi k A k ) 1 b k be the lossless function whose realization is R k. Then g k satisfies the interpolation condition Moreover, g k ( ) = γ k, γ k < 1. g k (z) = γ k z + g k 1 (z) z + γ k g k 1 (z) g k 1 (z) = (g k(z) γ k )z 1 γ k g k (z). This is the Schur algorithm: lossless functions of McMillan degree n are parametrized by the interpolation values γ n, γ n 1,..., γ 1 and γ 0. 13

14 An interpolation problem. Nevanlinna-Pick: find all the lossless functions G(z) G(1/ w)u = v, w < 1, u = 1. A solution exists v < 1. All the solutions are of the form G = (Θ 4 K + Θ 3 )(Θ 2 K + Θ 1 ) 1. for some lossless function K, where ( ( ) z w Θ(z) = I 2p + 1 wz 1 Θ = [ Θ1 Θ 2 Θ 3 Θ 4 ], J = [ I 0 0 I ], C = [ u ref. on interpolation pbs: Ball, Gohberg, Rodman, 1990 v ] ) C C J H (2), H cstarbit. H JH = J 14

15 Charts from a Schur algorithm. Schur algorithm: from G L p n, G = G n,..., G k LF T G k 1,..., G 0 G k (1/ w k )u k = v k, v k < 1 G k has degree k and G 0 is a constant unitary matrix. Chart and Schur parameters: w 1, w 2,..., w n points of the unit circle, u 1, u 2,..., u n unit complex p-vectors, v 1, v 2,..., v n complex p-vectors, v i < 1. W = {G L p n; G k (1/ w k )u k < 1}, ϕ : G (v 1,..., v n, G 0 ) 15

16 Balanced realizations from Schur parameters. For a particular choice of H in (2) depending on the interpolation data, a lossless function G in this chart has a unitary realization matrix computed by induction D k B k C k A k = V k 0 0 I k D k 1 C k 1 0 B k 1 A k 1 where A k is k k, D k is p p, and D 0 = G 0. very nice numerical behavior U k 0 0 I k 1, B. Hanzon, M.O., R. Peeters, INRIA report (2004) 16

17 The matrices U k and V k. U k and V k are the (p + 1) (p + 1) unitary matrices: U k = V k = ξ ku k η k w k ξ kv k η k I p (1 + η k w k )u k u k ξ k u k I p (1 η k ) v kv k v k 2 ξ k v k ξ k = 1 wk 2 1 wk 2 v k 2, η k = 1 vk 2 1 wk 2 v k 2 17

18 Adapted chart From a realization in Schur form (A lower triangular) A = w n w 1, B = 1 wn 2 u n by induction a chart in which v 1 = v 2 =... v n = 0 (Potapov factorization). Doesn t work for real functions: G( z) = G(z) Hanzon, Olivi, Peeters, ECC99 18

19 The RARL2 software This software computes a stable rational L 2 -approximation of specified order n to a multivariable transfer function given in one of the following forms: a realization a finite number of Fourier coefficients some pointwise values on the unit circle. It has been implemented using standard MATLAB subroutines. The optimizer of the toolkit OPTIM is used to find a local minimum, given by a realization, of the nonlinear L 2 -criterion. Implementation : J.P. Marmorat 19

20 Main features state-space representation parametrized by Schur parameters optimization over a manifold adapted chart obtained from a realization in Schur form recursive search on the degree (optional): minimum of degree k initial point of degree k + 1 (error-norm preserved) 20

21 Automobile gas turbine Hung, MacFarlane (1982) ; Glover (1984) ; Yan, Lam (1999) 21

22 Nyquist diagrams 2 2; order

23 Approximants: order 1 23

24 Approximants: order 2 24

25 Approximants: order 3 25

26 Approximants: order 4 26

27 Approximants: order 5 27

28 Approximants: order 6 28

29 Approximants: order 7 29

30 Approximants: order 8 30

31 Hyperfrequency Filter The problem: find a 8th order model of a MIMO (2 2) hyperfrequency filter, from experimental pointwise values in some range of frequencies provided by the CNES (French space agency). Data pre-processing (interpolation/completion): compute a stable matrix transfer function of high order which approximates these data, given by a great number (800) of Fourier coefficients. PRESTO-HF: software by F. Seyfert; HYPERION: software by J. Grimm. 31

32 Data and approximant at order 8 Nyquist diagrams 32

33 Data and approximant at order 8 Bode diagrams 33

34 1 2iπ Nudelman interpolation. T G(1/z)U (z I W ) 1 dz = V, (U, W ) output normal pair W W + U U = I (normalization) condition for a solution : positive definite solution to Lyapunov eq. P W P W = U U V V, P > 0. Then, G is the LFT of some K lossless G = (Θ 4 K + Θ 3 )(Θ 2 K + Θ 1 ) 1. Θ(z) = [ I 2p + (z 1)C(I zw ) 1 P 1 (I W ) C J ] H. Θ = [ Θ1 Θ 2 Θ 3 Θ 4 ], J = [ I 0 0 I Nudelman, 1977; Ball, Gohberg, Rodman, 1990 ], C = [ U V ] 34

35 Atlases of charts A chart: (W l, U l ),..., (W 1, U 1 ) output normal pairs, W k : n k n k, U k : p n k, k n k =n G L p n is in the chart if we can construct : G = G l,..., G k LF T G k 1,..., G 0. where G k satisfies Nudelman for (W k, U k, V k ) and P k > 0. The coordinate map: G (V 1,..., V l ) W k = w k, w k D, Nudelman G k (1/ w k )u k = v k, P k > 0 : v k < u k For a suitable choice of the LFT s, balanced realizations computed recursively from interpolation data (W k, U k, V k ). 35

36 Charts for real lossless functions W k, U k real, W k : 1 1 or 2 2 with two conjugate eighenvalues. Adapted chart : realization in real Schur form V 1 = V 2 =... V l = 0. W l 0 0 A =.... 0, B = U l. W 1 36

37 Allpass mutual encoding. W : n n 1-step Schur algorithm chart C Ω : (W, U) Ω(z) = Y + U(zI n W ) 1 X L p n P W P W = U U V V G(z) C Ω P > 0 (positive definite) adapted chart : Ω(z) = G(z), then P = I V = 0 Implementation : J.P. Marmorat 37

38 A finite atlas. A sub-atlas of Nevanlinna-Pick atlas (real case): Charts: w 1 = w 2 = = w n = 0; u 1, u 2,..., u n standard basis vectors. Balanced realization matrix= orthogonal positive p-upper Hessenberg matrix, up to a column permutation. Connection with the canonical form of (Hanzon, Ober, 1998) Peeters, Hanzon, M.O. 38

39 Hessenberg form. v n M n v n 1 M n M n 1 v n 2... M n M n 1... M 1 κ n vnv n 1 vnm n 1 v n 2... vnm n 1... M 1 0 κ n 1 vn 1v n 2... vn 1M n 2... M κ n 2 vn 2M n 3... M κ 2 v2v 1 v2m κ 1 v1, κ k = 1 v k 2, M k = I p + (κ k 1) v kv k v k 2. 39

40 Conclusion. parametrizations which present a nice behavior with respect to optimization problems an interesting connection between these parametrizations and realizations with particular structure a rich framework subclasses of functions or realizations with particular structure optimization parameters physical parameters olivi@sophia.inria.fr 40

Matrix Rational H 2 Approximation: a State-Space Approach using Schur Parameters

Matrix Rational H 2 Approximation: a State-Space Approach using Schur Parameters Matrix Rational H 2 Approximation: a State-Space Approach using Schur Parameters J.P. Marmorat, M. Olivi, B. Hanzon, R. Peeters Sophia Antipolis, France 1 Introduction We present a method to compute a

More information

Schur parametrizations and balanced realizations of real discrete-time stable all-pass systems.

Schur parametrizations and balanced realizations of real discrete-time stable all-pass systems. 1 Schur parametrizations and balanced realizations of real discrete-time stable all-pass systems Martine Olivi Jean-Paul Marmorat Bernard Hanzon Ralf LM Peeters Abstract We investigate the parametrization

More information

NUDELMAN INTERPOLATION, PARAMETRIZATIONS OF LOSSLESS FUNCTIONS AND BALANCED REALIZATIONS.

NUDELMAN INTERPOLATION, PARAMETRIZATIONS OF LOSSLESS FUNCTIONS AND BALANCED REALIZATIONS. NUDELMAN INTERPOLATION, PARAMETRIZATIONS OF LOSSLESS FUNCTIONS AND BALANCED REALIZATIONS. Jean-Paul Marmorat Martine Olivi CMA, BP 93, 692 Sophia-Antipolis Cedex, FRANCE, marmorat@sophia.inria.fr, phone:

More information

Balanced realization of lossless systems: Schur parameters, canonical forms and applications

Balanced realization of lossless systems: Schur parameters, canonical forms and applications Proceedings of the 15th IFAC Symposium on System Identification Balanced realization of lossless systems: Schur parameters, canonical forms and applications Ralf Peeters Martine Olivi Bernard Hanzon Department

More information

CANONICAL LOSSLESS STATE-SPACE SYSTEMS: STAIRCASE FORMS AND THE SCHUR ALGORITHM

CANONICAL LOSSLESS STATE-SPACE SYSTEMS: STAIRCASE FORMS AND THE SCHUR ALGORITHM CANONICAL LOSSLESS STATE-SPACE SYSTEMS: STAIRCASE FORMS AND THE SCHUR ALGORITHM Ralf L.M. Peeters Bernard Hanzon Martine Olivi Dept. Mathematics, Universiteit Maastricht, P.O. Box 616, 6200 MD Maastricht,

More information

Subdiagonal pivot structures and associated canonical forms under state isometries

Subdiagonal pivot structures and associated canonical forms under state isometries Preprints of the 15th IFAC Symposium on System Identification Saint-Malo, France, July 6-8, 29 Subdiagonal pivot structures and associated canonical forms under state isometries Bernard Hanzon Martine

More information

Canonical lossless state-space systems: staircase forms and the Schur algorithm

Canonical lossless state-space systems: staircase forms and the Schur algorithm Canonical lossless state-space systems: staircase forms and the Schur algorithm Ralf L.M. Peeters Bernard Hanzon Martine Olivi Dept. Mathematics School of Mathematical Sciences Projet APICS Universiteit

More information

Identification of microwave filters by analytic and rational H2 approximation

Identification of microwave filters by analytic and rational H2 approximation Identification of microwave filters by analytic and rational H2 approximation Martine Olivi, Fabien Seyfert, Jean-Paul Marmorat To cite this version: Martine Olivi, Fabien Seyfert, Jean-Paul Marmorat.

More information

On some interpolation problems

On some interpolation problems On some interpolation problems A. Gombani Gy. Michaletzky LADSEB-CNR Eötvös Loránd University Corso Stati Uniti 4 H-1111 Pázmány Péter sétány 1/C, 35127 Padova, Italy Computer and Automation Institute

More information

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

Schur functions. J. Rovnyak and H. S. V. de Snoo Schur functions J. Rovnyak and H. S. V. de Snoo The Schur class in complex analysis is the set of holomorphic functions S(z) which are defined and satisfy S(z) 1 on the unit disk D = {z : z < 1} in the

More information

OPTIMAL H 2 MODEL REDUCTION IN STATE-SPACE: A CASE STUDY

OPTIMAL H 2 MODEL REDUCTION IN STATE-SPACE: A CASE STUDY OPTIMAL H 2 MODEL REDUCTION IN STATE-SPACE: A CASE STUDY RLM Peeters, B Hanzon, D Jibetean Department of Mathematics, Universiteit Maastricht, PO Box 66, 6200 MD Maastricht, The Netherlands, E-mail: ralfpeeters@mathunimaasnl

More information

Introduction to The Dirichlet Space

Introduction to The Dirichlet Space Introduction to The Dirichlet Space MSRI Summer Graduate Workshop Richard Rochberg Washington University St, Louis MO, USA June 16, 2011 Rochberg () The Dirichlet Space June 16, 2011 1 / 21 Overview Study

More information

Definition We say that a topological manifold X is C p if there is an atlas such that the transition functions are C p.

Definition We say that a topological manifold X is C p if there is an atlas such that the transition functions are C p. 13. Riemann surfaces Definition 13.1. Let X be a topological space. We say that X is a topological manifold, if (1) X is Hausdorff, (2) X is 2nd countable (that is, there is a base for the topology which

More information

Boundary Nevanlinna Pick interpolation problems for generalized Schur functions

Boundary Nevanlinna Pick interpolation problems for generalized Schur functions , Boundary Nevanlinna Pick interpolation problems for generalized Schur functions Vladimir Bolotnikov and Alexander Kheifets Abstract. Three boundary Nevanlinna-Pick interpolation problems at finitely

More information

Comparing Two Generalized Nevanlinna-Pick Theorems

Comparing Two Generalized Nevanlinna-Pick Theorems Comparing Two Generalized Nevanlinna-Pick Theorems University of Iowa INFAS April 30, 2016 Comparing Two s Outline Motivation 1 Motivation 2 3 4 5 Comparing Two s Classical Nevanlinna-Pick Theorem Let

More information

Rational bases for system identification

Rational bases for system identification Rational bases for system identification Adhemar Bultheel, Patrick Van gucht Department Computer Science Numerical Approximation and Linear Algebra Group (NALAG) K.U.Leuven, Belgium adhemar.bultheel.cs.kuleuven.ac.be

More information

Subspace-based Identification

Subspace-based Identification of Infinite-dimensional Multivariable Systems from Frequency-response Data Department of Electrical and Electronics Engineering Anadolu University, Eskişehir, Turkey October 12, 2008 Outline 1 2 3 4 Noise-free

More information

5. Orthogonal matrices

5. Orthogonal matrices L Vandenberghe EE133A (Spring 2017) 5 Orthogonal matrices matrices with orthonormal columns orthogonal matrices tall matrices with orthonormal columns complex matrices with orthonormal columns 5-1 Orthonormal

More information

Parametric Nevanlinna-Pick Interpolation Theory

Parametric Nevanlinna-Pick Interpolation Theory Proc Natl Sci Counc ROC(A) Vol, No 6, 1998 pp 73-733 Parametric Nevanlinna-Pick Interpolation Theory FANG-BO YEH*, CHIEN-CHANG LIN**, AND HUANG-NAN HUANG* *Department of Mathematics Tunghai University

More information

H 2 -optimal model reduction of MIMO systems

H 2 -optimal model reduction of MIMO systems H 2 -optimal model reduction of MIMO systems P. Van Dooren K. A. Gallivan P.-A. Absil Abstract We consider the problem of approximating a p m rational transfer function Hs of high degree by another p m

More information

7.2 Conformal mappings

7.2 Conformal mappings 7.2 Conformal mappings Let f be an analytic function. At points where f (z) 0 such a map has the remarkable property that it is conformal. This means that angle is preserved (in the sense that any 2 smooth

More information

Corona Theorems for Multiplier Algebras on B n

Corona Theorems for Multiplier Algebras on B n Corona Theorems for Multiplier Algebras on B n Brett D. Wick Georgia Institute of Technology School of Mathematics Multivariate Operator Theory Banff International Research Station Banff, Canada August

More information

Stability of Parameter Adaptation Algorithms. Big picture

Stability of Parameter Adaptation Algorithms. Big picture ME5895, UConn, Fall 215 Prof. Xu Chen Big picture For ˆθ (k + 1) = ˆθ (k) + [correction term] we haven t talked about whether ˆθ(k) will converge to the true value θ if k. We haven t even talked about

More information

Berezin-Töplitz Quantization Math 242 Term Paper, Spring 1999 Ilan Hirshberg

Berezin-Töplitz Quantization Math 242 Term Paper, Spring 1999 Ilan Hirshberg 1 Introduction Berezin-Töplitz Quantization Math 242 Term Paper, Spring 1999 Ilan Hirshberg The general idea of quantization is to find a way to pass from the classical setting to the quantum one. In the

More information

Inverse source estimation problems in magnetostatics

Inverse source estimation problems in magnetostatics Inverse source estimation problems in magnetostatics ERNSI Workshop, Lyon, 2017 Juliette Leblond INRIA Sophia Antipolis (France), Team APICS Joint work with: Laurent Baratchart, Sylvain Chevillard, Doug

More information

Minimal symmetric Darlington synthesis

Minimal symmetric Darlington synthesis Minimal symmetric Darlington synthesis L. Baratchart, P. Enqvist, A. Gombani, M. Olivi November 9, 2007 Abstract. We consider the symmetric Darlington synthesis of a p p rational symmetric Schur function

More information

Adaptative Decomposition: The Case of the Drury Arveson Space

Adaptative Decomposition: The Case of the Drury Arveson Space DOI 10.1007/s00041-016-9508-4 Adaptative Decomposition: The Case of the Drury Arveson Space Daniel Alpay 1,2 Fabrizio Colombo 3 Tao Qian 4 Irene Sabadini 3 Received: 14 February 2016 Springer Science+Business

More information

FFTs in Graphics and Vision. Groups and Representations

FFTs in Graphics and Vision. Groups and Representations FFTs in Graphics and Vision Groups and Representations Outline Groups Representations Schur s Lemma Correlation Groups A group is a set of elements G with a binary operation (often denoted ) such that

More information

FEL3210 Multivariable Feedback Control

FEL3210 Multivariable Feedback Control FEL3210 Multivariable Feedback Control Lecture 5: Uncertainty and Robustness in SISO Systems [Ch.7-(8)] Elling W. Jacobsen, Automatic Control Lab, KTH Lecture 5:Uncertainty and Robustness () FEL3210 MIMO

More information

Iterative methods for positive definite linear systems with a complex shift

Iterative methods for positive definite linear systems with a complex shift Iterative methods for positive definite linear systems with a complex shift William McLean, University of New South Wales Vidar Thomée, Chalmers University November 4, 2011 Outline 1. Numerical solution

More information

The Takagi problem on the disk and bidisk

The Takagi problem on the disk and bidisk The Takagi problem on the disk and bidisk Jim Agler Joseph A. Ball John E. McCarthy December 11, 2012 Abstract We give a new proof on the disk that a Pick problem can be solved by a rational function that

More information

Learning gradients: prescriptive models

Learning gradients: prescriptive models Department of Statistical Science Institute for Genome Sciences & Policy Department of Computer Science Duke University May 11, 2007 Relevant papers Learning Coordinate Covariances via Gradients. Sayan

More information

A Parameterization of Positive Real Residue Interpolants with McMillan Degree Constraint

A Parameterization of Positive Real Residue Interpolants with McMillan Degree Constraint A Parameterization of Positive Real Residue Interpolants with McMillan Degree Constraint Yohei Kuroiwa Abstract A parameterization of the solutions to the positive real residue interpolation with McMillan

More information

Tensor algebras and subproduct systems arising from stochastic matrices

Tensor algebras and subproduct systems arising from stochastic matrices Tensor algebras and subproduct systems arising from stochastic matrices Daniel Markiewicz (Ben-Gurion Univ. of the Negev) Joint Work with Adam Dor-On (Univ. of Waterloo) OAOT 2014 at ISI, Bangalore Daniel

More information

Linear Passive Stationary Scattering Systems with Pontryagin State Spaces

Linear Passive Stationary Scattering Systems with Pontryagin State Spaces mn header will be provided by the publisher Linear Passive Stationary Scattering Systems with Pontryagin State Spaces D. Z. Arov 1, J. Rovnyak 2, and S. M. Saprikin 1 1 Department of Physics and Mathematics,

More information

Changing coordinates to adapt to a map of constant rank

Changing coordinates to adapt to a map of constant rank Introduction to Submanifolds Most manifolds of interest appear as submanifolds of others e.g. of R n. For instance S 2 is a submanifold of R 3. It can be obtained in two ways: 1 as the image of a map into

More information

Greedy Algorithms and Rational Approximation in One and Several Variables

Greedy Algorithms and Rational Approximation in One and Several Variables Greedy Algorithms and Rational Approximation in One and Several Variables Laurent Baratchart, Wei-Xiong Mai, Tao Qian To cite this version: Laurent Baratchart, Wei-Xiong Mai, Tao Qian. Greedy Algorithms

More information

Problem Set 5 Solutions 1

Problem Set 5 Solutions 1 Massachusetts Institute of Technology Department of Electrical Engineering and Computer Science 6.245: MULTIVARIABLE CONTROL SYSTEMS by A. Megretski Problem Set 5 Solutions The problem set deals with Hankel

More information

Introduction to Index Theory. Elmar Schrohe Institut für Analysis

Introduction to Index Theory. Elmar Schrohe Institut für Analysis Introduction to Index Theory Elmar Schrohe Institut für Analysis Basics Background In analysis and pde, you want to solve equations. In good cases: Linearize, end up with Au = f, where A L(E, F ) is a

More information

1 Hermitian symmetric spaces: examples and basic properties

1 Hermitian symmetric spaces: examples and basic properties Contents 1 Hermitian symmetric spaces: examples and basic properties 1 1.1 Almost complex manifolds............................................ 1 1.2 Hermitian manifolds................................................

More information

X t = a t + r t, (7.1)

X t = a t + r t, (7.1) Chapter 7 State Space Models 71 Introduction State Space models, developed over the past 10 20 years, are alternative models for time series They include both the ARIMA models of Chapters 3 6 and the Classical

More information

arxiv: v1 [math.cv] 24 Nov 2015

arxiv: v1 [math.cv] 24 Nov 2015 ADAPTIVE ORTHONORMAL SYSTEMS FOR MATRIX-VALUED FUNCTIONS DANIEL ALPAY, FABRIZIO COLOMBO, TAO QIAN, AND IRENE SABADINI arxiv:1511.07834v1 [math.cv] 4 Nov 015 Abstract. In this paper we consider functions

More information

Multivariable ARMA Systems Making a Polynomial Matrix Proper

Multivariable ARMA Systems Making a Polynomial Matrix Proper Technical Report TR2009/240 Multivariable ARMA Systems Making a Polynomial Matrix Proper Andrew P. Papliński Clayton School of Information Technology Monash University, Clayton 3800, Australia Andrew.Paplinski@infotech.monash.edu.au

More information

INTERPOLATION FOR MULTIPLIERS ON REPRODUCING KERNEL HILBERT SPACES

INTERPOLATION FOR MULTIPLIERS ON REPRODUCING KERNEL HILBERT SPACES PROCEEDINGS OF THE AMERICAN MATHEMATICAL SOCIETY Volume 131, Number 5, Pages 1373 1383 S 0002-9939(02)06899-5 Article electronically published on December 6, 2002 INTERPOLATION FOR MULTIPLIERS ON REPRODUCING

More information

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

A functional model for commuting pairs of contractions and the symmetrized bidisc A functional model for commuting pairs of contractions and the symmetrized bidisc Nicholas Young Leeds and Newcastle Universities Lecture 2 The symmetrized bidisc Γ and Γ-contractions St Petersburg, June

More information

Lecture 3: QR-Factorization

Lecture 3: QR-Factorization Lecture 3: QR-Factorization This lecture introduces the Gram Schmidt orthonormalization process and the associated QR-factorization of matrices It also outlines some applications of this factorization

More information

LUCK S THEOREM ALEX WRIGHT

LUCK S THEOREM ALEX WRIGHT LUCK S THEOREM ALEX WRIGHT Warning: These are the authors personal notes for a talk in a learning seminar (October 2015). There may be incorrect or misleading statements. Corrections welcome. 1. Convergence

More information

Math 108b: Notes on the Spectral Theorem

Math 108b: Notes on the Spectral Theorem Math 108b: Notes on the Spectral Theorem From section 6.3, we know that every linear operator T on a finite dimensional inner product space V has an adjoint. (T is defined as the unique linear operator

More information

Composition operators: the essential norm and norm-attaining

Composition operators: the essential norm and norm-attaining Composition operators: the essential norm and norm-attaining Mikael Lindström Department of Mathematical Sciences University of Oulu Valencia, April, 2011 The purpose of this talk is to first discuss the

More information

Smooth Dynamics 2. Problem Set Nr. 1. Instructor: Submitted by: Prof. Wilkinson Clark Butler. University of Chicago Winter 2013

Smooth Dynamics 2. Problem Set Nr. 1. Instructor: Submitted by: Prof. Wilkinson Clark Butler. University of Chicago Winter 2013 Smooth Dynamics 2 Problem Set Nr. 1 University of Chicago Winter 2013 Instructor: Submitted by: Prof. Wilkinson Clark Butler Problem 1 Let M be a Riemannian manifold with metric, and Levi-Civita connection.

More information

Discrete-time dichotomous well-posed linear systems and generalized Schur-Nevanlinna-Pick interpolation

Discrete-time dichotomous well-posed linear systems and generalized Schur-Nevanlinna-Pick interpolation Compl. anal. oper. theory 99 (9999), 154 1661-8254/99000-0, DOI 10.1007/s11785-003-0000 c 2006 Birkhäuser Verlag Basel/Switzerland Complex Analysis and Operator Theory Discrete-time dichotomous well-posed

More information

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

CONSERVATIVE DILATIONS OF DISSIPATIVE N-D SYSTEMS: THE COMMUTATIVE AND NON-COMMUTATIVE SETTINGS. Dmitry S. Kalyuzhnyĭ-Verbovetzkiĭ CONSERVATIVE DILATIONS OF DISSIPATIVE N-D SYSTEMS: THE COMMUTATIVE AND NON-COMMUTATIVE SETTINGS Dmitry S. Kalyuzhnyĭ-Verbovetzkiĭ Department of Mathematics Ben-Gurion University of the Negev P.O. Box 653

More information

Numerical Methods I Orthogonal Polynomials

Numerical Methods I Orthogonal Polynomials Numerical Methods I Orthogonal Polynomials Aleksandar Donev Courant Institute, NYU 1 donev@courant.nyu.edu 1 Course G63.2010.001 / G22.2420-001, Fall 2010 Nov. 4th and 11th, 2010 A. Donev (Courant Institute)

More information

Nonlinear palindromic eigenvalue problems and their numerical solution

Nonlinear palindromic eigenvalue problems and their numerical solution Nonlinear palindromic eigenvalue problems and their numerical solution TU Berlin DFG Research Center Institut für Mathematik MATHEON IN MEMORIAM RALPH BYERS Polynomial eigenvalue problems k P(λ) x = (

More information

Part II. Geometry and Groups. Year

Part II. Geometry and Groups. Year Part II Year 2014 2013 2012 2011 2010 2009 2008 2007 2006 2005 2014 Paper 4, Section I 3F 49 Define the limit set Λ(G) of a Kleinian group G. Assuming that G has no finite orbit in H 3 S 2, and that Λ(G),

More information

Model reduction for linear systems by balancing

Model reduction for linear systems by balancing Model reduction for linear systems by balancing Bart Besselink Jan C. Willems Center for Systems and Control Johann Bernoulli Institute for Mathematics and Computer Science University of Groningen, Groningen,

More information

A system coupling and Donoghue classes of Herglotz-Nevanlinna functions

A system coupling and Donoghue classes of Herglotz-Nevanlinna functions A system coupling and Donoghue classes of Herglotz-Nevanlinna functions S. Belyi, K. A. Makarov and E. Tsekanovskiĭ Abstract. We study the impedance functions of conservative L-systems with the unbounded

More information

MIMO Capacities : Eigenvalue Computation through Representation Theory

MIMO Capacities : Eigenvalue Computation through Representation Theory MIMO Capacities : Eigenvalue Computation through Representation Theory Jayanta Kumar Pal, Donald Richards SAMSI Multivariate distributions working group Outline 1 Introduction 2 MIMO working model 3 Eigenvalue

More information

Project-Team apics. Analysis and Problems of Inverse type in Control and Signal processing

Project-Team apics. Analysis and Problems of Inverse type in Control and Signal processing INSTITUT NATIONAL DE RECHERCHE EN INFORMATIQUE ET EN AUTOMATIQUE Project-Team apics Analysis and Problems of Inverse type in Control and Signal processing Sophia Antipolis - Méditerranée Theme : Modeling,

More information

Lecture III: Neighbourhoods

Lecture III: Neighbourhoods Lecture III: Neighbourhoods Jonathan Evans 7th October 2010 Jonathan Evans () Lecture III: Neighbourhoods 7th October 2010 1 / 18 Jonathan Evans () Lecture III: Neighbourhoods 7th October 2010 2 / 18 In

More information

Advances in Manifold Learning Presented by: Naku Nak l Verm r a June 10, 2008

Advances in Manifold Learning Presented by: Naku Nak l Verm r a June 10, 2008 Advances in Manifold Learning Presented by: Nakul Verma June 10, 008 Outline Motivation Manifolds Manifold Learning Random projection of manifolds for dimension reduction Introduction to random projections

More information

SCHUR-CLASS MULTIPLIERS ON THE FOCK SPACE: DE BRANGES-ROVNYAK REPRODUCING KERNEL SPACES AND TRANSFER-FUNCTION REALIZATIONS

SCHUR-CLASS MULTIPLIERS ON THE FOCK SPACE: DE BRANGES-ROVNYAK REPRODUCING KERNEL SPACES AND TRANSFER-FUNCTION REALIZATIONS SCHUR-CLASS MULTIPLIERS ON THE FOCK SPACE: DE BRANGES-ROVNYAK REPRODUCING KERNEL SPACES AND TRANSFER-FUNCTION REALIZATIONS JOSEPH A. BALL, VLADIMIR BOLOTNIKOV, AND QUANLEI FANG Abstract. We introduce and

More information

TRUNCATED TOEPLITZ OPERATORS ON FINITE DIMENSIONAL SPACES

TRUNCATED TOEPLITZ OPERATORS ON FINITE DIMENSIONAL SPACES TRUNCATED TOEPLITZ OPERATORS ON FINITE DIMENSIONAL SPACES JOSEPH A. CIMA, WILLIAM T. ROSS, AND WARREN R. WOGEN Abstract. In this paper, we study the matrix representations of compressions of Toeplitz operators

More information

1 Math 241A-B Homework Problem List for F2015 and W2016

1 Math 241A-B Homework Problem List for F2015 and W2016 1 Math 241A-B Homework Problem List for F2015 W2016 1.1 Homework 1. Due Wednesday, October 7, 2015 Notation 1.1 Let U be any set, g be a positive function on U, Y be a normed space. For any f : U Y let

More information

Project-Team Apics. Analysis and Problems of Inverse type in Control and Signal processing

Project-Team Apics. Analysis and Problems of Inverse type in Control and Signal processing INSTITUT NATIONAL DE RECHERCHE EN INFORMATIQUE ET EN AUTOMATIQUE Project-Team Apics Analysis and Problems of Inverse type in Control and Signal processing Sophia Antipolis THEME NUM d' ctivity eport 2006

More information

A Kupka-Smale Theorem for Hamiltonian systems. systems from Mañé s viewpoint, a control approach. Ludovic Rifford PICOF 12

A Kupka-Smale Theorem for Hamiltonian systems. systems from Mañé s viewpoint, a control approach. Ludovic Rifford PICOF 12 A Kupka-Smale Theorem for Hamiltonian systems from Mañé s viewpoint, a control approach Université de Nice - Sophia Antipolis & Institut Universitaire de France PICOF 12 The Kupka-Smale Theorem for vector

More information

Qualifying Exams I, Jan where µ is the Lebesgue measure on [0,1]. In this problems, all functions are assumed to be in L 1 [0,1].

Qualifying Exams I, Jan where µ is the Lebesgue measure on [0,1]. In this problems, all functions are assumed to be in L 1 [0,1]. Qualifying Exams I, Jan. 213 1. (Real Analysis) Suppose f j,j = 1,2,... and f are real functions on [,1]. Define f j f in measure if and only if for any ε > we have lim µ{x [,1] : f j(x) f(x) > ε} = j

More information

Finite-dimensional spaces. C n is the space of n-tuples x = (x 1,..., x n ) of complex numbers. It is a Hilbert space with the inner product

Finite-dimensional spaces. C n is the space of n-tuples x = (x 1,..., x n ) of complex numbers. It is a Hilbert space with the inner product Chapter 4 Hilbert Spaces 4.1 Inner Product Spaces Inner Product Space. A complex vector space E is called an inner product space (or a pre-hilbert space, or a unitary space) if there is a mapping (, )

More information

Composition Operators with Multivalent Symbol

Composition Operators with Multivalent Symbol Composition Operators with Multivalent Symbol Rebecca G. Wahl University of South Dakota, Vermillion, South Dakota 57069 March 10, 007 Abstract If ϕ is an analytic map of the unit disk D into itself, the

More information

11 COMPLEX ANALYSIS IN C. 1.1 Holomorphic Functions

11 COMPLEX ANALYSIS IN C. 1.1 Holomorphic Functions 11 COMPLEX ANALYSIS IN C 1.1 Holomorphic Functions A domain Ω in the complex plane C is a connected, open subset of C. Let z o Ω and f a map f : Ω C. We say that f is real differentiable at z o if there

More information

AN OPERATOR THEORETIC APPROACH TO DEGENERATED NEVANLINNA-PICK INTERPOLATION

AN OPERATOR THEORETIC APPROACH TO DEGENERATED NEVANLINNA-PICK INTERPOLATION 1 AN OPERATOR THEORETIC APPROACH TO DEGENERATED NEVANLINNA-PICK 1 Introduction INTERPOLATION Harald Woracek Institut für Technische Mathematik Technische Universität Wien A-1040 Wien, Austria In this paper

More information

Noise Reduction of JPEG Images by Sampled-Data H Optimal ε Filters

Noise Reduction of JPEG Images by Sampled-Data H Optimal ε Filters SICE Annual Conference 25 in Okayama, August 8-1, 25 Okayama University, Japan Noise Reduction of JPEG Images by Sampled-Data H Optimal ε Filters H. Kakemizu,1, M. Nagahara,2, A. Kobayashi,3, Y. Yamamoto,4

More information

The University of Texas at Austin Department of Electrical and Computer Engineering. EE381V: Large Scale Learning Spring 2013.

The University of Texas at Austin Department of Electrical and Computer Engineering. EE381V: Large Scale Learning Spring 2013. The University of Texas at Austin Department of Electrical and Computer Engineering EE381V: Large Scale Learning Spring 2013 Assignment Two Caramanis/Sanghavi Due: Tuesday, Feb. 19, 2013. Computational

More information

Inner Product Spaces An inner product on a complex linear space X is a function x y from X X C such that. (1) (2) (3) x x > 0 for x 0.

Inner Product Spaces An inner product on a complex linear space X is a function x y from X X C such that. (1) (2) (3) x x > 0 for x 0. Inner Product Spaces An inner product on a complex linear space X is a function x y from X X C such that (1) () () (4) x 1 + x y = x 1 y + x y y x = x y x αy = α x y x x > 0 for x 0 Consequently, (5) (6)

More information

Eigenvalue and Eigenvector Problems

Eigenvalue and Eigenvector Problems Eigenvalue and Eigenvector Problems An attempt to introduce eigenproblems Radu Trîmbiţaş Babeş-Bolyai University April 8, 2009 Radu Trîmbiţaş ( Babeş-Bolyai University) Eigenvalue and Eigenvector Problems

More information

Math 413/513 Chapter 6 (from Friedberg, Insel, & Spence)

Math 413/513 Chapter 6 (from Friedberg, Insel, & Spence) Math 413/513 Chapter 6 (from Friedberg, Insel, & Spence) David Glickenstein December 7, 2015 1 Inner product spaces In this chapter, we will only consider the elds R and C. De nition 1 Let V be a vector

More information

Lecture 2: Numerical linear algebra

Lecture 2: Numerical linear algebra Lecture 2: Numerical linear algebra QR factorization Eigenvalue decomposition Singular value decomposition Conditioning of a problem Floating point arithmetic and stability of an algorithm Linear algebra

More information

COMPLEX ANALYSIS AND RIEMANN SURFACES

COMPLEX ANALYSIS AND RIEMANN SURFACES COMPLEX ANALYSIS AND RIEMANN SURFACES KEATON QUINN 1 A review of complex analysis Preliminaries The complex numbers C are a 1-dimensional vector space over themselves and so a 2-dimensional vector space

More information

BERGMAN KERNEL ON COMPACT KÄHLER MANIFOLDS

BERGMAN KERNEL ON COMPACT KÄHLER MANIFOLDS BERGMAN KERNEL ON COMPACT KÄHLER MANIFOLDS SHOO SETO Abstract. These are the notes to an expository talk I plan to give at MGSC on Kähler Geometry aimed for beginning graduate students in hopes to motivate

More information

Chapter 6 Inner product spaces

Chapter 6 Inner product spaces Chapter 6 Inner product spaces 6.1 Inner products and norms Definition 1 Let V be a vector space over F. An inner product on V is a function, : V V F such that the following conditions hold. x+z,y = x,y

More information

These definitions show little of the nature of the spaces, however, and they are best understood through equivalent forms.

These definitions show little of the nature of the spaces, however, and they are best understood through equivalent forms. BULLETIN (New Series) OF THE AMERICAN MATHEMATICAL SOCIETY Volume 33, Number 1, January 1996 Sub-Hardy Hilbert spaces in the unit disk, by D. Sarason, Lecture Notes in the Mathematical Sciences, vol. 10,

More information

Computational Methods CMSC/AMSC/MAPL 460. Vectors, Matrices, Linear Systems, LU Decomposition, Ramani Duraiswami, Dept. of Computer Science

Computational Methods CMSC/AMSC/MAPL 460. Vectors, Matrices, Linear Systems, LU Decomposition, Ramani Duraiswami, Dept. of Computer Science Computational ethods CSC/ASC/APL 460 Vectors, atrices, Linear Systems, LU Decomposition, Ramani Duraiswami, Dept. of Computer Science Class Outline uch of scientific computation involves solution of linear

More information

Robust Control 2 Controllability, Observability & Transfer Functions

Robust Control 2 Controllability, Observability & Transfer Functions Robust Control 2 Controllability, Observability & Transfer Functions Harry G. Kwatny Department of Mechanical Engineering & Mechanics Drexel University /26/24 Outline Reachable Controllability Distinguishable

More information

Qualifying Examination HARVARD UNIVERSITY Department of Mathematics Tuesday, January 19, 2016 (Day 1)

Qualifying Examination HARVARD UNIVERSITY Department of Mathematics Tuesday, January 19, 2016 (Day 1) Qualifying Examination HARVARD UNIVERSITY Department of Mathematics Tuesday, January 19, 2016 (Day 1) PROBLEM 1 (DG) Let S denote the surface in R 3 where the coordinates (x, y, z) obey x 2 + y 2 = 1 +

More information

Matrix Lie groups. and their Lie algebras. Mahmood Alaghmandan. A project in fulfillment of the requirement for the Lie algebra course

Matrix Lie groups. and their Lie algebras. Mahmood Alaghmandan. A project in fulfillment of the requirement for the Lie algebra course Matrix Lie groups and their Lie algebras Mahmood Alaghmandan A project in fulfillment of the requirement for the Lie algebra course Department of Mathematics and Statistics University of Saskatchewan March

More information

Qualifying Exams I, 2014 Spring

Qualifying Exams I, 2014 Spring Qualifying Exams I, 2014 Spring 1. (Algebra) Let k = F q be a finite field with q elements. Count the number of monic irreducible polynomials of degree 12 over k. 2. (Algebraic Geometry) (a) Show that

More information

Linear Algebra Review

Linear Algebra Review Chapter 1 Linear Algebra Review It is assumed that you have had a course in linear algebra, and are familiar with matrix multiplication, eigenvectors, etc. I will review some of these terms here, but quite

More information

The Hodge Star Operator

The Hodge Star Operator The Hodge Star Operator Rich Schwartz April 22, 2015 1 Basic Definitions We ll start out by defining the Hodge star operator as a map from k (R n ) to n k (R n ). Here k (R n ) denotes the vector space

More information

Throughout these notes we assume V, W are finite dimensional inner product spaces over C.

Throughout these notes we assume V, W are finite dimensional inner product spaces over C. Math 342 - Linear Algebra II Notes Throughout these notes we assume V, W are finite dimensional inner product spaces over C 1 Upper Triangular Representation Proposition: Let T L(V ) There exists an orthonormal

More information

On Input Design for System Identification

On Input Design for System Identification On Input Design for System Identification Input Design Using Markov Chains CHIARA BRIGHENTI Masters Degree Project Stockholm, Sweden March 2009 XR-EE-RT 2009:002 Abstract When system identification methods

More information

Adaptive Filters. un [ ] yn [ ] w. yn n wun k. - Adaptive filter (FIR): yn n n w nun k. (1) Identification. Unknown System + (2) Inverse modeling

Adaptive Filters. un [ ] yn [ ] w. yn n wun k. - Adaptive filter (FIR): yn n n w nun k. (1) Identification. Unknown System + (2) Inverse modeling Adaptive Filters - Statistical digital signal processing: in many problems of interest, the signals exhibit some inherent variability plus additive noise we use probabilistic laws to model the statistical

More information

On the notions of dilation, controllability, observability, and minimality in the theory of dissipative scattering linear nd systems

On the notions of dilation, controllability, observability, and minimality in the theory of dissipative scattering linear nd systems On the notions of dilation, controllability, observability, and minimality in the theory of dissipative scattering linear nd systems Dmitriy S. Kalyuzhniy Department of Higher Mathematics Odessa State

More information

Geometric control and dynamical systems

Geometric control and dynamical systems Université de Nice - Sophia Antipolis & Institut Universitaire de France 9th AIMS International Conference on Dynamical Systems, Differential Equations and Applications Control of an inverted pendulum

More information

S p e c i a l M a t r i c e s. a l g o r i t h m. intert y msofdiou blystoc

S p e c i a l M a t r i c e s. a l g o r i t h m. intert y msofdiou blystoc M D O : 8 / M z G D z F zw z z P Dẹ O B M B N O U v O NO - v M v - v v v K M z z - v v MC : B ; 9 ; C vk C G D N C - V z N D v - v v z v W k W - v v z v O v : v O z k k k q k - v q v M z k k k M O k v

More information

Carathéodory-Julia type conditions and symmetries of boundary asymptotics for analytic functions on the unit disk

Carathéodory-Julia type conditions and symmetries of boundary asymptotics for analytic functions on the unit disk Carathéodory-Julia type conditions and symmetries of boundary asymptotics for analytic functions on the unit disk Vladimir Bolotnikov and Alexander Kheifets Mathematics Subject Classification (2000. 41A05,

More information

REPRESENTATION THEORY WEEK 7

REPRESENTATION THEORY WEEK 7 REPRESENTATION THEORY WEEK 7 1. Characters of L k and S n A character of an irreducible representation of L k is a polynomial function constant on every conjugacy class. Since the set of diagonalizable

More information

Topology III Home Assignment 8 Subhadip Chowdhury

Topology III Home Assignment 8 Subhadip Chowdhury Topology Home Assignment 8 Subhadip Chowdhury Problem 1 Suppose {X 1,..., X n } is an orthonormal basis of T x M for some point x M. Then by definition, the volume form Ω is defined to be the n form such

More information

Conformal field theory in the sense of Segal, modified for a supersymmetric context

Conformal field theory in the sense of Segal, modified for a supersymmetric context Conformal field theory in the sense of Segal, modified for a supersymmetric context Paul S Green January 27, 2014 1 Introduction In these notes, we will review and propose some revisions to the definition

More information

Practice Qualifying Exam Questions, Differentiable Manifolds, Fall, 2009.

Practice Qualifying Exam Questions, Differentiable Manifolds, Fall, 2009. Practice Qualifying Exam Questions, Differentiable Manifolds, Fall, 2009. Solutions (1) Let Γ be a discrete group acting on a manifold M. (a) Define what it means for Γ to act freely. Solution: Γ acts

More information