Zeros of Spectral Factors, the Geometry of Splitting Subspaces, and the Algebraic Riccati Inequality

In this paper it is shown how the zero dynamics of (not necessarily square) spectral factors relate to the splitting subspace geometry of stationary stochastic models and to the corresponding algebraic Riccati inequality. The notion of output-induced subspace of a minimal Markovian splitting subspace, which is the stochastic analogue of the supremal output-nulling subspace in geometric control theory, is introduced. Through this concept, the analysis can be made coordinate-free and straightforward geometric methods can be applied. It is shown how the zero structure of the family of spectral factors relates to the geometric structure of the family of minimal Markovian splitting subspaces in the sense that the relationship between the zeros of different spectral factors is reflected in the partial ordering of minimal splitting subspaces. Finally, the well-known characterization of the solutions of the algebraic Riccati equation is generalized in terms of Lagrangian subspaces invariant under the corresponding Hamiltonian to the larger solution set of the algebraic Riccati inequality.

[1]  A. MacFarlane.  An Eigenvector Solution of the Optimal Linear Regulator Problem , 1963 .

[2]  Enders A. Robinson.  Properties of the Wold Decomposition of Stationary Stochastic Processes , 1963 .

[3]  J. Potter.  Matrix Quadratic Solutions , 1966 .

[4]  P. Graefe.  Linear stochastic systems , 1966 .

[5]  B. Anderson.  The inverse problem of stationary covariance generation , 1969 .

[6]  G. Basile,et al.  Controlled and conditioned invariant subspaces in linear system theory , 1969 .

[7]  J. Willems.  Least squares stationary optimal control and the algebraic Riccati equation , 1971 .

[8]  K. Mårtensson,et al.  On the matrix riccati equation , 1971, Inf. Sci..

[9]  K. Mrtensson.  On the matrix riccati equation , 1971 .

[10]  Charles A. Desoer,et al.  Zeros and poles of matrix transfer functions and their dynamical interpretation , 1974 .

[11]  Harry Dym,et al.  Gaussian processes, function theory, and the inverse spectral problem , 1976 .

[12]  J. Pearson.  Linear multivariable control, a geometric approach , 1977 .

[13]  B.P. Molinari,et al.  The time-invariant linear-quadratic optimal control problem , 1977, Autom..

[14]  P. Caines,et al.  Splitting subspaces, spectral factorization and the positive real equation: Structural features of the stochastic realization problem , 1980, 1980 19th IEEE Conference on Decision and Control including the Symposium on Adaptive Processes.

[15]  M. Shayman.  Geometry of the Algebraic Riccati Equation, Part I , 1983 .

[16]  L. Silverman,et al.  System structure and singular control , 1983 .

[17]  A. Isidori,et al.  A frequency domain philosophy for nonlinear systems, with applications to stabilization and to adaptive control , 1984, The 23rd IEEE Conference on Decision and Control.

[18]  G. Picci,et al.  Realization Theory for Multivariate Stationary Gaussian Processes , 1985 .

[19]  Anders Lindquist,et al.  Forward and backward semimartingale models for gaussian processes with stationary increments , 1985 .

[20]  L. E. Faibusovich.  Matrix Riccati inequality: Existence of solutions , 1987 .

[21]  A. Isidori,et al.  On the nonlinear equivalent of the notion of transmission zeros , 1988 .

[22]  Michael Green,et al.  Balanced stochastic realizations , 1988 .

[23]  György Michaletzky,et al.  Zeros of (Non-Square) Spectral Factors and Canonical Correlations , 1990 .

[24]  C. Scherer.  The solution set of the algebraic Riccati equation and the algebraic Riccati inequality , 1991 .

[25]  Vladimír Kučera,et al.  Algebraic Riccati Equation: Hermitian and Definite Solutions , 1991 .

[26]  S. Bittanti,et al.  The Riccati equation , 1991 .