PARAMETER IDENTIFICATION PROBLEMS IN MATHEMATICAL MODELLING OF PROCESSES

Author(s): Avdeenko Tatiana Vladimirovna

Rubric: Information technology

Release: 2014-1 (4)

Pages: 115-124

Keywords: parameter identification, model structure, experimental design, local and global identifiability, parameter space

Annotation: The main stages of constructing mathematical models based on experimental data are consi-dered. At each stage the main problems of parametric identification leading to inadequate models are indicated. Special attention is paid to the property of model structure identifiability. Unidentified model structure contains a set of parameters that cannot be uniquely determined from experimental data, even for the ideal experiment. In this case the results of modeling become questionable. Illustrative examples of non-identifiable model structure are given, as well as recommendations for solving this problem.

Bibliography: Avdeenko TA.VL. PARAMETER IDENTIFICATION PROBLEMS IN MATHEMATICAL MODELLING OF PROCESSES // Education Resources and Technologies. – 2014. – № 1 (4). – С. 115-124. doi: