Application of Global Optimization Toolbox for Identification of Parameters of Interval Nonlinear Models of Static Systems

The problem of parametric identification of interval models of static systems as mathematically is an optimization problem with a set of solutions - interval estimates of parameters. When the models are nonlinear functions relative to parameters, it is almost impossible to obtain a set of solutions in the form of interval parameter values. The article is proposed that in such cases, instead of a set of parameters, only one vector to calculate and an interval model of the static system to build based on its. At the same time, the problems of parametric identification of interval nonlinear models of a static system are reduced to the problem of minimizing the mean square deviation between the results of modeling the characteristics of the object and the values that belong to the numerical intervals of this characteristic that obtained experimentally. In this way, a multidimensional optimization problem with a nonlinear multi-extremal objective function is obtained. To solve this problem, we're proposed using the standard tools of MATLAB Global Optimization Toolbox. The effectiveness of the existing tools of MATLAB global optimization for solving this problem is shown.