Conditions for the identifiability of mathematical models of evolutionary processes from the results of discrete indirect measurements of the state vector (Q1921552)

The problem of statistical solvability of problems of parametric identification of mathematical models of evolutionary processes with limited samples of experimental data (in the form of results of a finite number of indirect measurements of the state vectors) is studied. It is shown that, when there is no a priori information on the initial conditions (also on the boundary conditions, for models with distributed parameters), a correct solution of this problem is possible when the number of rows of the measurement matrix (which has full row rank) is not less than the dimension of the state vector.