DOCSLIB.ORG

A Sort-Jacobi Algorithm on Semisimple Lie Algebras∗ 1

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
A Sort-Jacobi Algorithm on Semisimple Lie Algebras∗ 1 A SORT-JACOBI ALGORITHM ON SEMISIMPLE LIE ALGEBRAS∗ M. KLEINSTEUBER† Abstract. A structure preserving Sort-Jacobi algorithm for computing eigenvalues or singular values is presented. The proposed method applies to an arbitrary semisimple Lie algebra on its ( 1)-eigenspace of the Cartan involution. Local quadratic convergence for arbitrary cyclic schemes is− shown for the regular case. The proposed method is independent of the representation of the underlying Lie algebra and generalizes well-known normal form problems such as the symmetric, Hermitian, skew-symmetric, symmetric and skew-symmetric R-Hamiltonian eigenvalue problem and the singular value decomposition. Due to the well-known classification of semisimple Lie algebras, the approach yields structure-preserving Jacobi eigenvalue methods for so far unstudied problems as Takagi’s and a Takagi-like factorization, the Hermitian quaternion, the Hermitian C-Hamiltonian eigenvalue decomposition and a symplectic singular value decomposition. The algorithm is exempli- fied for an exceptional Lie algebra. Key words. Jacobi algorithm, Cartan decomposition, semisimple Lie algebra, restricted-root space, structured eigenvalue decomposition, structured singular value decomposition, Takagi’s fac- torization, local quadratic convergence AMS subject classifications. 16D70, 45C05, 65F15, 15A18, 17B20, 74P20, 41A25 1. Introduction. Structured eigenvalue and singular value problems frequently arise in areas such as control theory and signal processing, where e.g. optimal control and signal estimation tasks require the fast computation of eigenvalues for Hamil- tonian and symmetric matrices, respectively. We refer to [4] for an extensive list of structured eigenvalue problems and relevant literature. Such problems cannot be solved easily by using standard software packages, as these eigenvalue algorithms do not necessarily preserve the underlying matrix structures and therefore may suffer by accumulation of rounding errors or numerical instabilities. In contrast, structure preserving eigenvalue methods have the potential of combining memory savings with high accuracy requirements and therefore are the methods of choice. Most of the previous approaches to structured problems have been on a case-to-case basis, by try- ing to adapt either known solution methods or software packages to each individual problem. However, such an ad–hoc approach is bound to fail if the problem of interest cannot be easily reformulated as a standard problem from numerical linear algebra. Thus a more flexible kind of algorithm is needed, that enables one to tackle whole classes of interesting eigenvalue problems. Clearly, one cannot expect to find one method that works for all structured problems. Nevertheless, it seems reasonable to develop methods that allow a unified approach to sufficiently rich classes of interesting eigenvalue problems. The structure preserving Jacobi-type methods proposed here are of this kind. Their inherent parallelizability, cf. [2], and high computational accuracy, cf. [11, 39] make them also useful for large scale matrix computations. They have moreover the potential to compete in convergence speed with state-of-the-art eigenvalue tools. Re- cently, a modified version of the one-sided Jacobi method outperforms the established QR-algorithm for computing the singular value decomposition of a dense matrix, cf. [9, 10]. ∗This paper contains results from the author’s PhD thesis, [34]. †National ICT Australia, Canberra Research Laboratory, SEACS Program, Locked Bag 8001, Canberra ACT 2601, Australia and Department of Information Engineering, RSISE, The Australian National University, Canberra, ACT 0200, Australia ([email protected]). 1 Variants of the Jacobi algorithm have been applied to many structured eigen- value problems (EVP), including e.g. the real skew-symmetric eigenvalue decompo- sition (EVD), [20, 31, 40], computations of the singular value decomposition (SVD) [36], non-symmetric EVPs [6, 14, 43, 46], complex symmetric EVPs [15], and the computation of eigenvalues of normal matrices [18]. For extensions to different types of generalized EVPs, we refer to [7, 47]. For applications of Jacobi–type methods to problems in systems theory, see [25, 26, 27]. In contrast to earlier work, we extend the classical concept of a Jacobi-algorithm towards a new, unified Lie algebraic approach to structured EVDs, where structure of a matrix is defined by being an element of a Lie algebra (or of a suitably defined sub-structure). To the best of the author’s knowledge, Wildberger [48] has been the first who pro- posed a generalization of the non-cyclic classical Jacobi algorithm on arbitrary com- pact Lie algebras and succeeded in proving global convergence of the algorithm. The well–known classification of compact Lie algebras shows that this approach essentially includes structure preserving Jacobi-type methods for the real skew-symmetric, the complex skew-Hermitian, the real skew-symmetric Hamiltonian, the complex skew- Hermitian Hamiltonian eigenvalue problem, and some exceptional cases. Wildberger’s work has been subsequently extended by Kleinsteuber et al. [33], where local quadratic convergence for general cyclic Jacobi schemes is shown to hold for arbitrary regular elements in a compact Lie algebra. In this paper we propose and analyze the Sort-Jacobi algorithm for a large class of structured matrices that, besides the compact Lie algebra cases, essentially includes the normal form problems quoted in Table 5. These cases arise from the well-known classification of simple Lie algebras. Cyclic Jacobi-type methods for some of these cases have been discussed earlier, e.g. for the symmetric/Hermitian EVD see [17, 31], for the skew-symmetric EVD see [20, 31, 40], for the real and complex SVD see [31, 36], for the real symmetric and skew-symmetric EVD see [16], for the Hermitian R-Hamiltonian EVD see [6], and for the perplectic EVD see [38]. Thus we achieve a simultaneous generalization of all this prior work. Note that the methods proposed in this paper exclusively use one-parameter transformations and therefore slightly differ from the algorithms in [6, 16, 20, 31, 40], where block Jacobi methods are used, i.e. multiparameter transformations that annihilate more than one pair of off-diagonal elements at the same time. Explicitly, our setting is the following. Let G be a semisimple Lie group and g = k p be the Cartan decomposition of its Lie algebra. We propose a Jacobi–type method⊕ that ”diagonalizes” an element S p by conjugation with some k K, where K G is the Lie subgroup corresponding∈ to k. Of course, if A is a Hermitian∈ matrix, then⊂ iA is skew-Hermitian and therefore the Hermitian and the skew-Hermitian EVD are equivalent. A similar trick extends to arbitrary compact Lie algebras. In this sense it becomes clear that the present work immediately includes the corresponding results for compact Lie algebras in [33]. In fact, if k is a compact semisimple Lie algebra, then g0 := k ik is the Cartan decomposition of g0. Thus the analysis of Jacobi-type methods on⊕ compact Lie algebras appears as a special case of our result. A characteristic feature of all known Jacobi-type methods is that they act to minimize the distance to diagonality while preserving the eigenvalues. Thus different measures to diagonality can be used to design different types of Jacobi algorithms. Conventional Jacobi algorithms, like classical cyclic Jacobi algorithms for symmet- ric matrices, Kogbetliantz’s method for the SVD, cf. [36], methods for the skew- symmetric EVD, cf. [20], and for the Hamiltonian EVD, cf. [37], as well as recent 2 methods for the perplectic EVD, cf. [38], are all based on reducing the sum of squares of off-diagonal entries (the so-called off-norm). Although local quadratic convergence to the diagonalization has been shown in some of these cases at least for generic situ- ations, the analysis of such conventional Jacobi methods becomes considerably more complicate for clustered eigenvalues. This difficulty is unavoidable for conventional Jacobi methods and is due to the fact, that the off-norm function that is to be min- imized has a complicated critical point structure and several global minima. Thus this difficulty might be remedied by a better choice of cost function that measures the distance to diagonality. In fact, Brockett’s trace function turns out to be a more ap- propriate distance measure than the off-norm function. In [3], R.W. Brockett showed that the gradient flow of the trace function can be used to diagonalize a symmetric matrix and simultaneously sort the eigenvalues. This trace function has also been considered by e.g. M.T. Chu, [8] associated with gradient methods for matrix factor- izations, and subsequently by many others. For a systematic critical point analysis of the trace function in a Lie group setting, we refer to [13]. See also [44] for more recent results on this topic in the framework of reductive Lie groups. In his PhD thesis [31], K. H¨uper has been the first who realized, that Brockett’s trace function can be effectively used to design a new kind of Jacobi algorithm for symmetric matrix diagonalization, that automatically sorts the eigenvalues in any prescribed order. This Sort-Jacobi algorithm uses Givens rotations that do not only annihilate the off-diagonal element, but also sort the two corresponding elements on the diagonal. The idea of combining sorting with eigenvalue computations can be carried over to the SVD as well, cf. [31]. It is this sorting property that distinguishes the Sort-Jacobi algorithm from the known conventional schemes and leads both to improved convergence properties as well as to a simplified theory. Lie theory provides us both with a unified treatment and a coordinate free ap- proach of Jacobi-type methods. In particular, it allows a formulation of Jacobi meth- ods that is independent of the underlying matrix representation. Together with the above specific matrix cases, all isomorphic types can be simultaneously treated. Thus we can completely avoid the often tiring case-by-case analysis. We illustrate this process in Subsection 5.2.
Load more

Recommended publications
  • 1 Introduction
    Real and Complex Hamiltonian Square Ro ots of SkewHamiltonian Matrices y z HeikeFab ender D Steven Mackey Niloufer Mackey x Hongguo Xu January DedicatedtoProfessor Ludwig Elsner on the occasion of his th birthday Abstract We present a constructive existence pro of that every real skewHamiltonian matrix W has a real Hamiltonian square ro ot The key step in this construction shows how one may bring anysuch W into a real quasiJordan canonical form via symplectic similarityWe show further that every W has innitely many real Hamiltonian square ro ots and givealower b ound on the dimension of the set of all such square ro ots Extensions to complex matrices are also presented This report is an updated version of the paper Hamiltonian squareroots of skew Hamiltonian matrices that appearedinLinear Algebra its Applicationsv pp AMS sub ject classication A A A F B Intro duction Any matrix X suchthatX A is said to b e a square ro ot of the matrix AF or general nn complex matrices A C there exists a welldevelop ed although somewhat complicated theory of matrix square ro ots and a numb er of algorithms for their eective computation Similarly for the theory and computation of real square ro ots for real matrices By contrast structured squareroot problems where b oth the matrix A Fachb ereich Mathematik und Informatik Zentrum fur Technomathematik Universitat Bremen D Bremen FRG email heikemathunibremende y Department of Mathematics Computer Science Kalamazo o College Academy Street Kalamazo o MI USA z Department of Mathematics
    [Show full text]
  • Eigenvalue Perturbation Theory of Structured Real Matrices and Their Sign Characteristics Under Generic Structured Rank-One Perturbations∗
    Eigenvalue perturbation theory of structured real matrices and their sign characteristics under generic structured rank-one perturbations∗ Christian Mehlx Volker Mehrmannx Andr´eC. M. Ran{ Leiba Rodmank July 9, 2014 Abstract An eigenvalue perturbation theory under rank-one perturbations is developed for classes of real matrices that are symmetric with respect to a non-degenerate bilinear form, or Hamiltonian with respect to a non-degenerate skew-symmetric form. In contrast to the case of complex matrices, the sign characteristic is a cru- cial feature of matrices in these classes. The behavior of the sign characteristic under generic rank-one perturbations is analyzed in each of these two classes of matrices. Partial results are presented, but some questions remain open. Appli- cations include boundedness and robust boundedness for solutions of structured systems of linear differential equations with respect to general perturbations as well as with respect to structured rank perturbations of the coefficients. Key Words: real J-Hamiltonian matrices, real H-symmetric matrices, indefinite inner product, perturbation analysis, generic perturbation, rank-one perturbation, T - even matrix polynomial, symmetric matrix polynomial, bounded solution of differential equations, robustly bounded solution of differential equations, invariant Lagrangian subspaces. Mathematics Subject Classification: 15A63, 15A21, 15A54, 15B57. xInstitut f¨urMathematik, TU Berlin, Straße des 17. Juni 136, D-10623 Berlin, Germany. Email: fmehl,[email protected]. {Afdeling Wiskunde, Faculteit der Exacte Wetenschappen, Vrije Universiteit Amsterdam, De Boele- laan 1081a, 1081 HV Amsterdam, The Netherlands and Unit for BMI, North West University, Potchef- stroom, South Africa. E-mail: [email protected] kCollege of William and Mary, Department of Mathematics, P.O.Box 8795, Williamsburg, VA 23187-8795, USA.
    [Show full text]
  • Hamiltonian Matrix Representations for the Determination of Approximate Wave Functions for Molecular Resonances
    Journal of Applied and Fundamental Sciences HAMILTONIAN MATRIX REPRESENTATIONS FOR THE DETERMINATION OF APPROXIMATE WAVE FUNCTIONS FOR MOLECULAR RESONANCES Robert J. Buenker Fachbereich C-Mathematik und Naturwissenschaften, Bergische Universität Wuppertal, Gaussstr. 20, D-42097 Wuppertal, Germany *For correspondence. ([email protected]) Abstract: Wave functions obtained using a standard complex Hamiltonian matrix diagonalization procedure are square integrable and therefore constitute only approximations to the corresponding resonance solutions of the Schrödinger equation. The nature of this approximation is investigated by means of explicit calculations using the above method which employ accurate diabatic potentials of the B 1Σ+ - D’ 1Σ+ vibronic resonance states of the CO molecule. It is shown that expanding the basis of complex harmonic oscillator functions gradually improves the description of the exact resonance wave functions out to ever larger internuclear distances before they take on their unwanted bound-state characteristics. The justification of the above matrix method has been based on a theorem that states that the eigenvalues of a complex-scaled Hamiltonian H (ReiΘ) are associated with the energy position and linewidth of resonance states (R is an internuclear coordinate and Θ is a real number). It is well known, however, that the results of the approximate method can be obtained directly using the unscaled Hamiltonian H (R) in real coordinates provided a particular rule is followed for the evaluation of the corresponding matrix elements. It is shown that the latter rule can itself be justified by carrying out the complex diagonalization of the Hamiltonian in real space via a product of two transformation matrices, one of which is unitary and the other is complex orthogonal, in which case only the symmetric scalar product is actually used in the evaluation of all matrix elements.
    [Show full text]
  • Hamiltonian Matrices and the Algebraic Riccati Equation Seminar Presentation
    Hamiltonian Matrices and the Algebraic Riccati Equation Seminar presentation by Piyapong Yuantong 7th December 2009 Technische Universit¨atChemnitz Piyapong Yuantong Hamiltonian Matrices and the Algebraic Riccati Equation 1 Hamiltonian matrices We start by introducing the square matrix J 2 R 2n×2n defined by " # O I J = n n ; (1) −In On where n×n On 2 R − zero matrix n×n In 2 R − identity matrix: Remark It is not very complicated to prove the following properties of the matrix J: i. J T = −J ii. J −1 = J T T iii. J J = I2n T T iv. J J = −I2n 2 v. J = −I2n vi. det J = 1 Definition 1 A matrix A 2 R 2n×2n is called Hamiltonian if JA is symmetric, so JA = (JA) T ) A T J + JA = 0 where J 2 R 2n×2n is from (1). We will denote { } Hn = A 2 R 2n×2n j A T J + JA = 0 the set of 2n × 2n Hamiltonian matrices. 1 Piyapong Yuantong Hamiltonian Matrices and the Algebraic Riccati Equation Proposition 1.1 The following are equivalent: a) A is a Hamiltionian matrix b) A = JS, where S = S T c) (JA) T = JA Proof a ! b A = JJ −1 A ! A = J(−J)A 2Hn A ! (J(−JA)) T J + JA = 0 ! (−JA) T J T J = −JA T J J=!I2n (−JA) T = −JA ! J(−JA) T = A If − JA = S ! A = J(−JA) = J(−JA) T ! A = JS = JS T =) S = S T a ! c A T J + JA = 0 ! A T J = −JA ()T ! (A T J) T = (−JA) T ! J T A = −(JA) T ! −JA = −(JA) T ! (JA) T = JA 2 Piyapong Yuantong Hamiltonian Matrices and the Algebraic Riccati Equation Proposition 1.2 Let A; B 2 Hn.
    [Show full text]
  • Mathematical Physics-14-Eigenvalue Problems.Nb
    Eigenvalue problems Main idea and formulation in the linear algebra The word "eigenvalue" stems from the German word "Eigenwert" that can be translated into English as "Its own value" or "Inherent value". This is a value of a parameter in the equation or system of equations for which this equation has a nontriv- ial (nonzero) solution. Mathematically, the simplest formulation of the eigenvalue problem is in the linear algebra. For a given square matrix A one has to find such values of l, for which the equation (actually the system of linear equations) A.X λX (1) has a nontrivial solution for a vector (column) X. Moving the right part to the left, one obtains the equation A − λI .X 0, H L where I is the identity matrix having all diagonal elements one and nondiagonal elements zero. This matrix equation has nontrivial solutions only if its determinant is zero, Det A − λI 0. @ D This is equivalent to a Nth order algebraic equation for l, where N is the rank of the mathrix A. Thus there are N different eigenvalues ln (that can be complex), for which one can find the corresponding eigenvectors Xn. Eigenvectors are defined up to an arbitrary numerical factor, so that usually they are normalized by requiring T∗ Xn .Xn 1, where X T* is the row transposed and complex conjugate to the column X. It can be proven that eigenvectors that belong to different eigenvalues are orthogonal, so that, more generally than above, one has T∗ Xm .Xn δmn . Here dmn is the Kronecker symbol, 1, m n δ = mn µ 0, m ≠ n.
    [Show full text]
  • MAT TRIAD 2019 Book of Abstracts
    MAT TRIAD 2019 International Conference on Matrix Analysis and its Applications Book of Abstracts September 8 13, 2019 Liblice, Czech Republic MAT TRIAD 2019 is organized and supported by MAT TRIAD 2019 Edited by Jan Bok, Computer Science Institute of Charles University, Prague David Hartman, Institute of Computer Science, Czech Academy of Sciences, Prague Milan Hladík, Department of Applied Mathematics, Charles University, Prague Miroslav Rozloºník, Institute of Mathematics, Czech Academy of Sciences, Prague Published as IUUK-ITI Series 2019-676ø by Institute for Theoretical Computer Science, Faculty of Mathematics and Physics, Charles University Malostranské nám. 25, 118 00 Prague 1, Czech Republic Published by MATFYZPRESS, Publishing House of the Faculty of Mathematics and Physics, Charles University in Prague Sokolovská 83, 186 75 Prague 8, Czech Republic Cover art c J. Na£eradský, J. Ne²et°il c Jan Bok, David Hartman, Milan Hladík, Miroslav Rozloºník (eds.) c MATFYZPRESS, Publishing House of the Faculty of Mathematics and Physics, Charles University, Prague, Czech Republic, 2019 i Preface This volume contains the Book of abstracts of the 8th International Conference on Matrix Anal- ysis and its Applications, MAT TRIAD 2019. The MATTRIAD conferences represent a platform for researchers in a variety of aspects of matrix analysis and its interdisciplinary applications to meet and share interests and ideas. The conference topics include matrix and operator theory and computation, spectral problems, applications of linear algebra in statistics, statistical models, matrices and graphs as well as combinatorial matrix theory and others. The goal of this event is to encourage further growth of matrix analysis research including its possible extension to other elds and domains.
    [Show full text]
  • Matrix G-Strands
    Technological University Dublin ARROW@TU Dublin Articles School of Mathematics 2014 Matrix G-strands Darryl Holm Imperial College London Rossen Ivanov Technological University Dublin, [email protected] Follow this and additional works at: https://arrow.tudublin.ie/scschmatart Part of the Mathematics Commons, Non-linear Dynamics Commons, and the Partial Differential Equations Commons Recommended Citation Holm, D. and Ivanov, R. Matrix G-Strands, Nonlinearity 27 (2014) 1445–1469; doi:10.1088/0951-7715/27/ 6/1445 This Article is brought to you for free and open access by the School of Mathematics at ARROW@TU Dublin. It has been accepted for inclusion in Articles by an authorized administrator of ARROW@TU Dublin. For more information, please contact [email protected], [email protected]. This work is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 4.0 License Funder: ERC Matrix G-Strands Darryl D. Holm1 and Rossen I. Ivanov2 9 Feb 2014, clean copy Abstract We discuss three examples in which one may extend integrable Euler–Poincar´e ODEs to integrable Euler–Poincar´ePDEs in the matrix G-Strand context. After de- scribing matrix G-Strand examples for SO(3) and SO(4) we turn our attention to SE(3) where the matrix G-Strand equations recover the exact rod theory in the con- vective representation. We then find a zero curvature representation (ZCR) of these equations and establish the conditions under which they are completely integrable. Thus, the G-Strand equations turn out to be a rich source of integrable systems. The treatment is meant to be expository and most concepts are explained in examples in the language of vectors in R3.
    [Show full text]
  • A Unified Approach to Linear and Nonlinear Normal Forms For
    View metadata, citation and similar papers at core.ac.uk brought to you by CORE provided by Elsevier - Publisher Connector J. Symbolic Computation (1999) 27, 49–131 Article No. jsco.1998.0244 Available online at http://www.idealibrary.com on A Unied Approach to Linear and Nonlinear Normal Forms for Hamiltonian Systems† R. C. CHURCHILL‡‡§AND MARTIN KUMMER§ ‡Department of Mathematics, Hunter College and Graduate Center, CUNY, 695 Park Avenue, NY 10021, U.S.A. § Department of Mathematics, University of Toldeo, Toledo, OH 43606, U.S.A. We prove that in all but one case the normal form of a real or complex Hamiltonian matrix which is irreducible and appropriately normalized can be computed by Lie series methods in formally the same manner as one computes the normal form of a nonlinear Hamiltonian function. Calculations are emphasized; the methods are illustrated with detailed examples, and for the sake of completeness the exceptional case is also reviewed and illustrated. Alternate methods are also discussed, along with detailed examples. c 1999 Academic Press 1. Introduction A linear operator T : V → V on a real nite-dimensional symplectic space V =(V,ω)is Hamiltonian,orinnitesimally symplectic,if ω(Tu,v)+ω(u, T v)=0,u,v∈ V. (1.1) The linear Hamiltonian normal form problem is that of nding simple matrix represen- tations for such operators relative to symplectic bases of V . The theoretical solution was provided by Williamson (1936) (also see Wall (1963)), but calculation of normal forms following his ideas proves dicult in practice. Simplications have since appeared, but the methods are still quite involved, (see, e.g.
    [Show full text]
  • Matrix Nearness Problems and Applications∗
    Matrix Nearness Problems and Applications∗ Nicholas J. Higham† Abstract A matrix nearness problem consists of finding, for an arbitrary matrix A, a nearest member of some given class of matrices, where distance is measured in a matrix norm. A survey of nearness problems is given, with particular emphasis on the fundamental properties of symmetry, positive definiteness, orthogonality, normality, rank-deficiency and instability. Theoretical results and computational methods are described. Applications of nearness problems in areas including control theory, numerical analysis and statistics are outlined. Key words. matrix nearness problem, matrix approximation, symmetry, positive definiteness, orthogonality, normality, rank-deficiency, instability. AMS subject classifications. Primary 65F30, 15A57. 1 Introduction Consider the distance function (1.1) d(A) = min E : A + E S has property P ,A S, k k ∈ ∈ where S denotes Cm×n or Rm×n, is a matrix norm on S, and P is a matrix property k · k which defines a subspace or compact subset of S (so that d(A) is well-defined). Associated with (1.1) are the following tasks, which we describe collectively as a matrix nearness problem: Determine an explicit formula for d(A), or a useful characterisation. • ∗This is a reprint of the paper: N. J. Higham. Matrix nearness problems and applications. In M. J. C. Gover and S. Barnett, editors, Applications of Matrix Theory, pages 1–27. Oxford University Press, 1989. †Department of Mathematics, University of Manchester, Manchester, M13 9PL, England ([email protected]). 1 Determine X = A + E , where E is a matrix for which the minimum in (1.1) • min min is attained.
    [Show full text]
  • H = Hamiltonian Matrix S = Overlap Matrix
    chmy564-19 Lec 8 3. Linear Variation Method Boot Camp Mon 28jan19 Hij = Hamiltonian matrix Sij = Overlap matrix Linear Variation with 2 orthonormal, real, basis functions Ψ = c1 Φ1 + c2 Φ2 (H11 − E) c1 + H12 c2 = 0 H21c2 + (H22 − E) c2 = 0 where H12 = <Φ1|H|Φ2>, the "interaction" of the two basis functions, H11 = <Φ1|H|Φ1>, the "energy" of Φ1, H22 = <Φ2|H|Φ2>, the "energy" of Φ2, and E = <Ψ|H|Ψ>. the expectation value of the energy using the trial function. For a non-trivial solution (c1 and c2 not both zero), the determinant of the matrix of numbers multiplying the coefficients c1 and c2 must vanish, i.e., (H11 −E) (H22 − E) − H21 H12 = 0 quadratic equation; Solve for E 2 E – E(H11 + H22) +(H11H22 − H21 H12)= 0 quadratic 2 2 −±b b −4 ac ()()4(HH11+± 22 HH11 +− 22 HHHH11 22 − 21 12 E± = = 22a 2 H11+H22 H11−H22 2 E± = ± +H Because H12 is real 2 2 12 H H α1 β H= 11 12 = β α H21 H22 2 2 (αα+− )( αα ) 2 E =±+12 12 β 22 Solving for Ψ = c1 Φ1 + c2 Φ2 means solving for c1 and c2 (H11 – E)c1 + (H12)c2 = 0 (α1 – E)c1 + βc2 = 0 (H21)c1 + (H22 – E)c2 = 0 = βc1 + (α2 – E)c2 = 0 The determinant from the N simultaneous homogeneous equations = 0 is a constraint that means that you can only solve for the ratio of the coefficients. 5 (α1 – E)c1 + βc2 = 0 The determinant from the N simultaneous homogeneous equations = 0 is a constraint that βc1 + (α2 – E)c2 = 0 means that you can only solve for the ratio of the coefficients.
    [Show full text]
  • Arxiv:2011.08087V2 [Math-Ph] 30 Mar 2021 [36, 6] of a Haar Measure Unitary
    THE GENERALIZED CARTAN DECOMPOSITION FOR CLASSICAL RANDOM MATRIX ENSEMBLES ALAN EDELMAN AND SUNGWOO JEONG Abstract. We present a completed classification of the classical random ma- trix ensembles: Hermite (Gaussian), Laguerre (Wishart), Jacobi (MANOVA) and Circular by introducing the concept of the generalized Cartan decompo- sition and the double coset space. Previous authors [8, 40] associate a symmetric space G=K with a random matrix density on the double coset structure KnG=K. However this is incom- plete. Complete coverage requires the double coset structure A = K1nG=K2, where G=K1 and G=K2 are two symmetric spaces. Furthermore, we show how the matrix factorization obtained by the generalized Cartan decomposition G = K1AK2 plays a crucial role in sampling algorithms and the derivation of the joint probability density of A. 1. Introduction In pioneering work by Zirnbauer [40] and Due~nez[8] connections are made be- tween random matrix ensembles and the symmetric spaces [19, 21] studied in dif- ferential geometry, representation theory, and harmonic analysis. Unfortunately, symmetric spaces alone do not account for a rich enough set of Jacobi ensemble densities (See Figure 1). This work fills this significant gap. Noting that more general tools are needed, we not only take a closer look at the KAK decomposition from the theory of symmetric spaces, but we propose the incorporation of the more recent generalized Cartan decomposition (or the K1AK2 decomposition) [15, 24, 14, 30] especially for complete coverage of the Jacobi ensembles. We show that double coset spaces are rich enough to cover these important matrix ensembles. Furthermore, the key object is not the joint density, per se, but the matrix factorization.
    [Show full text]
  • Technische Universit at Chemnitz
    Technische Universitat Chemnitz Sonderforschungsbereich Numerische Simulation auf massiv parallelen Rechnern Peter Benner Ralph Byers Volker Mehrmann Hongguo Xu Numerical Computation of Deating Subspaces of Embedded Hamiltonian Pencils Preprint SFB PreprintReihe des Chemnitzer SFB SFB June Contents Introduction and Preliminaries Embedding of Extended Matrix Pencils Hamiltonian Triangular Forms SkewHamiltonianHamiltonian Matrix Pencils Algorithms Error and Perturbation Analysis Conclusion A Algorithmic Details A Schur Forms for SkewHamiltonianHamiltonian Matrix Pencils and SkewHa miltonianskewHamiltonian Matrix Pencils A Subroutines Required by Algorithm Authors Peter Benner Ralph Byers Zentrum f ur Technomathematik Department of Mathematics Fachbereich Mathematik und Informatik University of Kansas Universitat Bremen Lawrence KS D Bremen FRG USA bennermathunibremende byersmathukansedu Volker Mehrmann Hongguo Xu Deparment of Mathematics Fakultat f ur Mathematik Case Western University Technische Universitat Chemnitz Cleveland OH D Chemnitz FRG USA mehrmannmathematiktuchemnitzde hxxpocwruedu Numerical Computation of Deating Subspaces of Embedded Hamiltonian Pencils 2 3 Peter Benner Ralph Byers Volker Mehrmann Hongguo Xu Abstract We discuss the numerical solution of structured generalized eigenvalue problems that arise from linearquadratic optimal control problems H optimization multibo dy sys 1 tems and many other areas of applied mathematics physics and chemistry The classical
    [Show full text]