Application of Genetic Algorithms to Vector Optimization of the Automatic Control Systems
Valeriy Petrovich Severin · eNTUKhPIIR Repository (Kharkiv Polytechnic Institute) · 2009
Methods for calculation of quality indexes for automatic control systems are presented. For the optimization of quality indexes defined only in a stability domain a vector objective function of varied parameters of the system is proposed. The stepwise principle of successive satisfaction of constraints for the passage into the definition domain of quality indexes is considered, as well as a rational mechanism of its realization in the form of the priority optimization of the vector objective function. For the optimization of the vector objective functions genetic algorithms as vector optimization methods are presented. Their application allows one to steer the optimization process from any initial point of the space of varied parameters into the stability domain of the system and to find the optimum of the quality indexes in this domain. The efficiency of the proposed application of vector genetic algorithms for the quality indexes optimization is confirmed by computational experiments.