MEOPT software package for solving non-convex problems of parametric identification

Pavel S. Sorokovikov, Alexander Yu. Gornov

Matrosov Institute for System Dynamics and Control Theory of SB RAS

This paper considers the MEOPT software package created by the authors for the numerical study of non-convex parametric identification tasks. The developed software includes libraries of optimization algorithms and test problems, tool and service modules, and metacomponents. The implemented libraries of optimization algorithms include methods of multidimensional and one-dimensional non-convex optimization. The software was created in the C language using the GCC compiler and supports operation on Windows, Linux and MacOS operating systems. To date, the prototypes of the main modules of the package have been completed. Performed technical testing of the first version of the software.

software package, parametric identification, dynamic models, optimization of dynamic systems