Report on test matrices for generalized inverses

This paper is a comprehensive report on test matrices for the generalized inversion of matrices. Two principles are described how to construct singular square or arbitrary rectangular test matrices and their Moore-Penrose inverses. By prescribing the singular values of the matrices or by suitably choosing the free parameters test matrices with condition numbers of any size can be obtained. We also deal with test matrices which are equal to their Moore-Penrose inverse. In addition to many advices how to construct test matrices the paper presents many test matrices explicitly, in particular singular square matrices of ordern, sets of 7×6 and 7×5 matrices of different rank, a set of 5×5 matrices which are equal to their Moore-Penrose inverse and some special test matrices known from literature. For the set of 7×6 parameter matrices also the singular values corresponding to six values of the parameter are listed. For three simple parameter matrices of order 5×4 and 6×5 even test results obtained by eight different algorithms are quoted.As “by-products” the paper contains inequalities between condition numbers of different norms, representations for unitary, orthogonal, column-orthogonal and row-orthogonal matrices, a generalization of Hadamard matrices and representations of matrices which are equal to their Moore-Penrose inverse (or their inverse). All test matrices given in this paper may also be used for testing algorithms solving linear least squares problems.ZusammenfassungDiese Arbeit stellt einen umfassenden Bericht über Testmatrizen für die verallgemeinerte Matrizeninversion dar. Es werden zwei Prinzipien beschrieben, wie man singuläre quadratische oder beliebige rechteckige Testmatrizen und ihre Moore-Penrose-Inversen konstruieren kann. Durch die Vorgabe der singulären Werte der Matrizen oder durch geeignete Wahl der freien Parameter können die Testmatrizen beliebig große Konditionszahlen annehmen. Es wird auch auf Testmatrizen eingegangen, die gleich ihrer Moore-Penrose-Inversen sind. Neben vielen Hinweisen, wie man selbst Testmatrizen konstruieren kann, werden in der Arbeit zahlreiche Testmatrizen explizit angegeben, insbesondere singuläre quadratische Matrizenn-ter Ordnung, Serien von (7×6)- und (7×5)-Matrizen verschiedenen Ranges, eine Serie von (5×5)-Matrizen, die gleich ihrer Moore-Penrose-Inversen sind und einige spezielle Testmatrizen aus der Literatur. Für die Serie der (7×6)-Parametermatrizen sind die singulären Werte für sechs Werte des Parameters mit angegeben. Für drei einfache Parametermatrizen vom Format 5×4 und 6×5 sind auch die mit acht verschiedenen Algorithmen erzielten Testergebnisse aufgeführt.Als „Nebenergebnisse” enthält die Arbeit Ungleichungen zwischen Konditionszahlen unterschiedlicher Normen, Darstellungen für unitäre, orthogonale, spaltenorthogonale und zeilenorthogonale Matrizen, eine Verallgemeinerung der Hadamard-Matrizen und Darstellungen von Matrizen, die gleich ihrer Moore-Penrose-Inversen (oder ihrer Inversen) sind. Die in der Arbeit aufgeführten Testmatrizen können auch zum Testen von Algorithmen zur Lösung von linearen Kleinste-Quadrate-Problemen verwendet werden.

[1]  Charles L. Lawson,et al.  Solving least squares problems , 1976, Classics in applied mathematics.

[2]  A. Berman.  Nonnegative matrices which are equal to their generalized inverse , 1974 .

[3]  James Hardy Wilkinson,et al.  Rounding errors in algebraic processes , 1964, IFIP Congress.

[4]  J. Neumann,et al.  Numerical inverting of matrices of high order , 1947 .

[5]  Adi Ben-Israel,et al.  Generalized inverses: theory and applications , 1974 .

[6]  M. Saunders,et al.  A Bidiagonalization Algorithm for Sparse Linear Equations and Least-Squares Problems. , 1978 .

[7]  Zahari Zlatev,et al.  A testing scheme for subroutines solving large linear problems , 1981, Comput. Chem..

[8]  W. Givens.  Numerical Computation of the Characteristic Values of a Real Symmetric Matrix , 1954 .

[9]  Gerhard Zielke.  Testmatrizen mit freien Parametern , 2005, Computing.

[10]  G. Stewart,et al.  On the Numerical Properties of an Iterative Method for Computing the Moore–Penrose Generalized Inverse , 1974 .

[11]  J. Westlake.  Handbook of Numerical Matrix Inversion and Solution of Linear Equations , 1968 .

[12]  Roy H. Wampler.  A Report on the Accuracy of Some Widely Used Least Squares Computer Programs , 1970 .

[13]  Gerhard Zielke,et al.  Testmatrizen mit maximaler Konditionszahl , 1974, Computing.

[14]  J. H. Wilkinson,et al.  Handbook for Automatic Computation. Vol II, Linear Algebra , 1973 .

[15]  G. Zielke.  Generalizations of a Rutishauser test matrix with exact Moore-Penrose inverses , 1981, SGNM.

[16]  C. D. Meyer,et al.  Generalized inverses of linear transformations , 1979 .

[17]  R. Penrose.  A Generalized inverse for matrices , 1955 .

[18]  John R. Rice.  SQUARS: AN ALGORITHM FOR LEAST-SQUARES APPROXIMATION**This work was supported in part by NSF Grant GP-05850. , 1971 .

[19]  Werner Sautter,et al.  Fehleranalyse für die Gauß-Elimination zur Berechnung der Lösung minimaler Länge , 1978 .

[20]  J. Neumann,et al.  Numerical inverting of matrices of high order. II , 1951 .

[21]  H. Heinrich.  R. Bellman, Introduction to Matrix Analysis. XX + 328 S. London 1960. McGraw-Hill. Preis geb. 77s. 6d , 1961 .

[22]  George D. Poole,et al.  A class of Hessenberg matrices with known pseudoinverse and Drazin inverse , 1975 .

[23]  J. H. Wilkinson.  Modern Error Analysis , 1971 .

[24]  Kern O. Kymn,et al.  Alternative computing formulas for the generalized inverse and the evaluation of their performances , 1979 .

[25]  Alston S. Householder,et al.  The Theory of Matrices in Numerical Analysis , 1964 .

[26]  C. Lawson,et al.  Solving least squares problems , 1976, Classics in applied mathematics.

[27]  Robert Todd Gregory,et al.  A collection of matrices for testing computational algorithms , 1969 .

[28]  Peter Deuflhard,et al.  On rank-deficient pseudoinverses , 1980 .

[29]  N. Abdelmalek.  Round off error analysis for Gram-Schmidt method and solution of linear least squares problems , 1971 .

[30]  L. Mirsky,et al.  Introduction to Linear Algebra , 1965, The Mathematical Gazette.

[31]  Gene H. Golub,et al.  Calculating the singular values and pseudo-inverse of a matrix , 2007, Milestones in Matrix Computation.

[32]  G. Stewart.  On the Perturbation of Pseudo-Inverses, Projections and Linear Least Squares Problems , 1977 .

[33]  Marilyn Bohl,et al.  Information processing , 1971 .

[34]  Nai–Kuan Tsao.  A Note on Implementing the Householder Transformation , 1975 .

[35]  Masaaki Sibuya,et al.  Numerical algorithms for the Moore-Penrose inverse of a matrix: Direct methods , 1972 .

[36]  H. Rutishauser.  Once again: The least square problem , 1968 .

[37]  G. Forsythe,et al.  Computer solution of linear algebraic systems , 1969 .

[38]  Roy H. Wampler.  An evaluation of linear least squares computer programs , 1969 .

[39]  James Hardy Wilkinson,et al.  Error Analysis of Direct Methods of Matrix Inversion , 1961, JACM.

[40]  Michael A. Saunders,et al.  LSQR: An Algorithm for Sparse Linear Equations and Sparse Least Squares , 1982, TOMS.

[41]  K. Tanabe,et al.  Conjugate-gradient method for computing the Moore-Penrose inverse and rank of a matrix , 1977 .

[42]  Basic Numerical Mathematics , 1977 .

[43]  B. Noble,et al.  Methods for Computing the Moorse-Penrose Generalized Inverse, and Related Matters , 1976 .

[44]  Richard Bellman,et al.  Introduction to Matrix Analysis , 1972 .

[45]  Ulkich Tautenhahn.  Zur Auswahl glatter Lösungen über der Lösungsmannigfaltigkeit unterbestimmter schlechtkonditionierter Systeme , 1981 .

[46]  Gene H. Golub,et al.  Singular value decomposition and least squares solutions , 1970, Milestones in Matrix Computation.

[47]  A. Sluis,et al.  Stability of the solutions of linear least squares problems , 1974 .

[48]  J. H. Wilkinson.  The algebraic eigenvalue problem , 1966 .

[49]  Jörn Springer.  Exact solution of general integer systems of linear equations , 1986, TOMS.

[50]  G. Zielke.  A survey of generalized matrix inverses , 1984 .

[51]  R. Penrose.  On best approximate solutions of linear matrix equations , 1956, Mathematical Proceedings of the Cambridge Philosophical Society.

[52]  T. L. Boullion,et al.  Computation of Pseudoinverse Matrices Using Residue Arithmetic , 1972 .

[53]  Å. Björck.  Solving linear least squares problems by Gram-Schmidt orthogonalization , 1967 .

[54]  G. Zielke.  Verallgemeinerungen einer Testmatrix von Rutishauser mit exakten MOORE-PENROSE-Inversen , 1981 .

[55]  D. Faddeev,et al.  Computational Methods of Linear Algebra , 1959 .

[56]  Morris Newman,et al.  The Evaluation of Matrix Inversion Programs , 1958 .

[57]  J. Springer.  Die exakte Berechnung der Moore-Penrose-Inversen einer Matrix durch Residuenarithmetik , 1983 .

[58]  D. Faddeev,et al.  Computational methods of linear algebra , 1959 .

[59]  P. Wedin.  Perturbation theory for pseudo-inverses , 1973 .

[60]  J. Uhlig.  C. Forsythe and C. B. Moler, Computer Solution of Linear Algebraic Systems. (Series in Automatic Computation) XI + 148 S. Englewood Cliffs, N.J. 1967. Prentice-Hall, Inc. Preis geb. 54 s. net , 1972 .

[61]  Nabih N. Abdelmalek,et al.  On the solution of the linear least squares problems and pseudo-inverses , 1974, Computing.

[62]  G. Stewart,et al.  Rank degeneracy and least squares problems , 1976 .

[63]  James W. Longley.  An Appraisal of Least Squares Programs for the Electronic Computer from the Point of View of the User , 1967 .

[64]  G. Golub,et al.  Iterative refinements of linear least squares solutions by Householder transformations , 1968 .

[65]  Ben Noble.  A Method for Computing the Generalized Inverse of a Matrix , 1966 .

[66]  Test matrices for generalized inverses , 1978, SGNM.

[67]  E. V. Krishnamurthy,et al.  Residue Arithmetic Algorithms for Exact Computation of g-Inverses of Matrices , 1976 .

[68]  Alston S. Householder,et al.  Handbook for Automatic Computation , 1960, Comput. J..

[69]  G. Stewart.  The Efficient Generation of Random Orthogonal Matrices with an Application to Condition Estimators , 1980 .

[70]  J. B. Rosen,et al.  Computation of the pseudoinverse of a matrix of unknown rank , 1964 .