IDENTIFIABILITY OF POLYNOMIAL SYSTEMS: STRUCTURAL AND NUMERICAL ASPECTS

The problem of obtaining meaningful estimates for the parameters of polynominal models is studied. A finite algebraic condition is given for deterministic identiflability of a class of polynomial models in zero-state experiments. Some open problems are formulated. The effects of sampling constraints and measurement noise on practical identiflability are taken into account in a method of analysis, based on eigenvalue — eigenvector decomposition of the approximate Hessian matrix of the least squares objective function. By uncovering dependences among the parameters, the method suggests simplifying assumptions in order to formulate a well-posed estimation problem.

[1]  K. Glover,et al.  Identifiability of linear and nonlinear dynamical systems , 1976 .

[2]  A. Krener,et al.  The existence and uniqueness of volterra series for nonlinear systems , 1977 .

[3]  J. Baillieul.  Controllability and observability of polynomial dynamical systems , 1981 .

[4]  John A. Jacquez,et al.  The inverse problem for compartmental systems , 1982 .

[5]  Y. Lecourtier,et al.  Algebraic Manipulation Routines for Testing Structural Properties , 1985 .

[6]  D. Allara,et al.  Parameterization of complex reaction systems: Model fitting vs. fundamental kinetics , 1973 .

[7]  J. Jacquez,et al.  Numerical parameter identifiability and estimability: Integrating identifiability, estimability, and optimal sampling design , 1985 .

[8]  H. Pohjanpalo.  System identifiability based on the power series expansion of the solution , 1978 .

[9]  A. Krener,et al.  Nonlinear controllability and observability , 1977 .

[10]  M. Fliess.  A note on volterra series expansions for nonlinear differential systems , 1980 .

[11]  B. Eisenberg,et al.  Kinetics of methane pyrolysis , 1967 .

[12]  Françoise Lamnabhi-Lagarrigue,et al.  An algebraic approach to nonlinear functional expansions , 1983 .

[13]  J. DiStefano,et al.  Optimal nonuniform sampling interval and test-input design for identification of physiological systems from very limited data , 1979 .

[14]  S. Vajda,et al.  Structural identifiability of dynamical systems , 1983 .

[15]  J. DiStefano,et al.  Identifiability of Model Parameter , 1985 .

[16]  Peter P. Valko,et al.  Principal component analysis of kinetic models , 1985 .

[17]  T. Daubert,et al.  Application of quasilinearization to methane pyrolysis , 1974 .

[18]  S. Vajda,et al.  Structural equivalence of linear systems and compartmental models , 1981 .