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

Personal portfolio

CONTROL SYSTEM PROBLEM SOLUTION BY VARIATIONAL GENETIC PROGRAMMING METHOD

Diveev A.I. 1 Ibadulla S.I. 2 Sofronova E.A. 2
1 Institution of Russian Academy of Science Dorodnicyn Computing Centre of RAS
2 Peoples’ Friendship University of Russia
We examine the problem of synthesis of control systems, where we need to find the control as a function of the space coordinates of the object’s state. To solve the problem it is proposed to use a new method of variational genetic programming. A description of the method of variational genetic programming is given. In contrast to the classical method of genetic programming in a new variational method for genetic programming all genetic operations are performed on sets of vectors describing small variations of possible solutions. Small variations in genetic programming are defined and a data structure as an integer vector to describe a small variation is proposed. To describe a possible solution there proposed to use an ordered set of vectors of the two components, the first one of which indicates the number of arguments to a function, and the second one indicates the function index. To describe a small variation a three components’ vector there used. The first component indicates the index of variation, the second part sets the points of variation, and the third component indicates the function index, if it is necessary while realizing the variation. There is a numerical example of the synthesis of mobile robot controlling system under spatial constraints conditions.