Scientific journal
Scientific Review. Technical science
ISSN 2500-0799
ПИ №ФС77-57440

Personal portfolio

GRAMMATICAL EVOLUTION AND NETWORK OPERATOR METHODS FOR SYNTHESIS OF THE CONTROL SYSTEM FOR A DYNAMIC OBJECT

Diveev A.I. 1 Kazaryan D.E. 2
1 Institution of Russian Academy of Sciences Dorodnicyn Computing Centre of RAS
2 Cybernetics and Mechatronics dept.
We considered the control system synthesis problem for the nonlinear dynamic object. The control synthesis problem is a problem of the search for the function of the object state. Synthesized system should reach the chosen goal for the given set of the initial conditions. We stated the goal as a set of a quality functionals. We used grammatical evolution method and network operator method to solve the given problem. Grammatical evolution is a subclass of the genetic programming field, that uses production rules system to build the mathematical expression. In the network operator method mathematical expression is represented as a graph. We chose the steady-state genetic algorithm as a search engine for both methods. The solution was presented in a form of the Pareto set, that contained a set of the satisfactory control functions. We chose the nonlinear Duffing oscillator as a dynamic object. We conducted he computational experiment. Both methods were proved to be able to solve the control system synthesis problem.