Abstract
The identification problem for linear stochastic systems may be stated roughly as follows: given observations on two stochastic processes which are the input and output of some unknown linear system, determine some estimate of the parameters of the system. A set of candidate linear systems which contains the “true” system is introduced, and probabilistic assumptions on the two stochastic processes turn the identification problem into the deterministic problem of minimizing some objective function over this candidate model set. If this set is a manifold, the existence of globally convergent identification algorithms hinges on the critical point behavior of the objective functions which it carries. By way of Morse Theory, the critical point behavior of objective functions on a manifold has implications with regard to the topology of the manifold. This paper analyzes the topology and critical point behavior of objective functions on a specific manifold of linear systems which appears frequently as the candidate model set in identification problems. This manifold is the set Σ of allm-input,p-output linear systems of fixed McMillan degree with real or complex coefficients. Over this manifold sits the principal bundle\(\tilde \Sigma\) of minimal realizations of systems in Σ It is shown that there exist three natural analytic metrics on the associated vector bundle. It is also shown that, in the real case, the first Stiefel-Whitney class of the bundle\(\tilde \Sigma\) has min(m, p)-1 nonvanishing powers; the same conclusion is drawn about the first Chern class of\(\tilde \Sigma\) in the complex case. These results, which follow from Morse Theory and some elementary homotopy and homology theory, imply that the category of the bundle\(\tilde \Sigma\) is at least min(m, p), and hence that the Lusternik-Schnirelmann category of Σ is at least min(m, p). It follows that canonical forms (i.e. sections of\(\tilde \Sigma\)) may exist only when min(m, p) = 1 and that any objective function on Σ with compact sublevel sets has at least min(m, p) critical points. In particular, there exist on Σ no globally convergent gradient algorithms when min(m, p) > 1.
Similar content being viewed by others
References
Astrom, K. and T. Soderstrom, Uniqueness of the maximum likelihood estimate of the parameters of an ARMA model,IEEE Transactions on Automatic Control, AC-19 (1974), pp. 769–773.
Bott, R.Lectures on Morse Theory. Bonn: Mathematische Institut der Universitat Bonn, 1960.
Bott, R. Morse-theoretic aspects of Yang-Mills theory, fromRecent Developments in Gauge Theory, G. 't Hooft, C. Itzvkson, A. Jaffe, H. Lehmann, P. K. Mitter, I. M. Singer, and R. Stora, Editors. London: Plenum Publishing Co., 1980.
Bott, R. The stable homotopy of the classical groups,Annals of Mathematics, 70 (1959), pp. 313–337.
Bott, R. and S. Chern, Hermitian vector bundles and the equidistribution of the zeroes of their holomorphic sections,Acta Mathematica, 114 (1965), pp. 71–112.
Brockett, R. W.Finite-Dimensional Linear Systems. New York: John Wiley and Sons, 1970.
Brockett, R. W., Some geometric questions in the theory of linear systems,IEEE Transactions on Automatic Control, AC-21 (1976), pp. 449–455.
Brockett, R. W., Structural properties of the equilibrium solutions to Riccati equations, inLecture Notes on Mathematics Volume 132. New York: Springer-Verlag, 1970, pp. 61–69.
Brockett, R. W. and P. S. Krishnaprasad, A scaling theory for linear systems,IEEE Transactions on Automatic Control, AC-25 (1980), pp. 197–207.
Byrnes, C. I., Algebraic and geometric aspects of the analysis of feedback systems, inGeometrical Methods for the Theory of Linear Systems, C. I. Brynes and C. F. Martin, editors. Boston: D. Reidel Publishing Co., 1980, pp. 85–124.
Byrnes, C. I., On the stabilizability of linear control systems depending on parameters,Proceedings of the 18th IEEE Conference on Decision and Control, Ft. Lauderdale, FL, Dec., 1979, pp. 233–236.
Byrnes, C. I., The moduli space for linear dynamical systems, inGeometric Control Theory, C. Martin and R. Hermann, editors. Brookline, MA: Math. Sci. Press. 1977, pp. 229–276.
Byrnes, C. I. and T. E. Duncan, On certain topological invariants arising in system theory, fromNew Directions in Applied Mathematics, P. Hilton and G. Young, editors. New York: Springer-Verlag, 1981, pp. 29–71.
Byrnes, C. I. and N. Hurt, On the moduli of linear dynamical systems,Advances in Mathematical Studies in Analysis, 4 (1979), pp. 83–122.
Caines, P. E., Prediction error identification methods for stationary stochastic processes,IEEE Transactions on Automatic Control, AC-21 (1976), pp. 500–505.
Caines, P. E., Stationary linear and nonlinear system identification and predictor set completeness,IEEE Transactions on Automatic Control, AC-23 (1978), pp. 583–594.
Clark, J. M. C., The consistent selection of local coordinates in linear system identification,Proceedings of the Joint Automatic Control Conference (1976), pp. 576–580.
Conner, P. E., Concerning the action of a finite group,Proceedings of the National Academy of Sciences, 41 (1956), pp. 349–351.
Delchamps, D. F., A note on the analyticity of the Riccati metric, inLectures on Applied Mathematics, Volume 18, C. I. Byrnes and C. F. Martin, editors, Providence, RI: American Mathematical Society, 1980, pp. 37–41.
Delchamps, David F., Analytic feedback control and the algebraic Riccati equation, to appear inIEEE Transactions on Automatic Control, AC-29 (1984).
Delchamps, David F., Analytic stabilization and the algebraic Riccati equation,Proceedings of the 22nd IEEE Conference on Decision and Control, San Antonio, TX, 1983, pp. 1396–1401.
Delchamps, David F., On families of linear systems: The fundamental group and the uniqueness of objective function minima, in preparation forSystems and Control Letters.
Delchamps, David F., Smooth input-output compensators for linear systems,Proceedings of the Eighteenth Annual Conference on Information Science and Systems, Princeton, NJ, 1984.
Delchamps, David F., The geometry of spaces of linear systems with an application to the identification problem, ph.D. Thesis, Harvard University, 1982.
Delchamps, David F. and C. I. Byrnes, Critical point behavior of objective functions defined on spaces of multivariable systems,Proceedings of the 21st IEEE Conference on Decision and Control, Orlando, FL, 1982, pp. 937–943.
Desoer, C. A.Notes for a Second Course on Linear Systems. New York: Van Nostrand Rienhold, 1970.
Diendonne, J.Foundations of Modern Analysis. New York: Academic Press, 1969.
Eilenberg, S. Sur une théorème topologique de M. L. Schnirelmann,Math Sb., 1 (1936), pp. 557–559.
Gantmacher, L.The Theory of Matrices. New York: Chelsea, 1959.
Glover, K.Structural Aspects of System Identification, Ph.D. Thesis, Department of Electrical Engineering, Electronic Systems Laboratory, Massachusetts Institute of Technology, Cambridge, MA, Report No. ESL-R-516.
Greenberg, M. and J. Harper.Algebraic Topology: A First Course. Reading, MA: The Benjamin/Cummings Publishing Co., 1981.
Griffiths, P. and J. Adams,Topics in Algebraic and Analytic Geometry. Princeton: Princeton University Press, 1974.
Guest, M.Some Relationships Between Homotopy Theory and Differential Geometry, Ph.D. Thesis, Wolfson College, University of Oxford, 1981.
Guillemin, V. and A. PollackDifferential Topology. Englewood Cliffs, NJ: Prentice Hall, 1974.
Hazewinkel, M. Moduli and canonical forms for linear dynamical systems II: The topological case,Math. Syst. Theor., 10 (1977), pp. 363–385.
Hazewinkel, M. and R. E. Kalman, On invariants, canonical forms, and moduli for linear constant, finite-dimensional dynamical systems, inLecture Notes in Econ.-Mathematical System Theory Volume 131. New York: Springer-Verlag, 1976, pp. 48–60.
Hazewinkel, M. and J. C. Willems, editors.Stochastic Systems: The Mathematics of Filtering and Identification and Applications. Dordrecht, Holland: D. Reidel Publishing Co., 1981.
Helmke, Uwe. Zur Topologie des raumes linearear Kontrollsysteme, Ph.D. Thesis, Universitat Bremen, 1982.
Hermann, R. and C. F. Martin, Applications of algebraic geometry to systems theory: The McMillan degree and Kronecker indices of transfer functions as topological and holomorphic system invariants,SIAM Journal on Control, 16 (1978), pp. 743–755.
Hirsch, M. W.Differential Topology. New York: Springer-Verlag, 1976.
James, I. M., On category, In the sense of Lusternik-Schnirelmann,Topology, 17 (1978), pp. 341–348.
Kalman, R. E., Algebraic geometric description of the class of linear systems of constant dimension, 8th Annual Conference on Information Sciences and Systems, Princeton, NJ, March, 1974.
Kalman, R. E., Mathematical description of linear dynamical systems,SIAM Journal on Control, 1 (1963), pp. 128–151.
Massey, W. S.Algebraic Topology: An Introduction. New York: Harcourt, Brace and World, 1971.
Massey, W. S.Singular Homology Theory. New York: Springer-Verlag, 1980.
Mehra, R. K. and D. G. Laniotis, editors.Systems Identification: Advances and Case Studies. New York: Academic Press, 1976.
Milnor, J. W.Morse Theory. Princeton: Princeton University Press, 1963.
Milnor, J. W.Topology From the Differentiable Viewpoint. Charlottesville, VA: University of Virginia Press, 1965.
Milnor, J. W. and J. D. Stasheff.Characteristic Classes. Princeton: Princeton University Press, 1974.
Nomizu, K.Lie Groups and Differential Geometry. Tokyo: Mathematical Society of Japan, 1956.
Segal, G., The topology of spaces of rational functions,Acta Mathematica, 143 (1979), pp. 39–72.
Spanier, E.Algebraic Topology. New York: McGraw Hill Book Co., 1966.
Spivak, M.A Comprehensive Introduction to Differential Geometry, Volumes I and II. Berkeley, CA: Publish or Perish, Inc. 1979.
Spivak, M.Calculus on Manifolds. Reading, MA: Benjamin, 1965.
Steenrod, N.The Topology of Fibre Bundles. Princeton: Princeton University Press, 1951.
Sussmann, H. J., and E. D. Sonntag Reachability does not imply pole assignability,Notices of the American Mathematical Society.
Willems, J. C., Least squares stationary optimal control and the algebraic Riccati equation,IEEE Transactions on Automatic Control, AC 16 (1971), pp. 621–634.
Wilson, F. W., The structure of the level surfaces of a Lyapunov function,Journal of Differential Equations, 3 (1967), pp. 323–329.
Wong, E.Stochastic Processes in Information and Dynamical Systems. New York: McGraw Hill Book Co., 1971.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Delchamps, D.F. Global structure of families of multivariable linear systems with an application to identification. Math. Systems Theory 18, 329–380 (1985). https://doi.org/10.1007/BF01699476
Received:
Revised:
Issue Date:
DOI: https://doi.org/10.1007/BF01699476