Bella L. Khashper, Olga G. Kantor

Ufa State Petroleum Technical University

The article discusses the application of random search method for finding the extremal value of a function that depends on multiple parameters. The advantages and disadvantages of the method are described, as well as the possibility of improving it using machine learning methods. Using the example of quantitative analysis of multicomponent mixtures, it is shown how random search method can be combined with machine learning methods to solve the problem of parametric identification.

Multidimensional optimization, random search method, statistical methods, machine learning, classification task, regression task, decision tree, Trust-region method, condition number