UDC 519.712.2

Authors

Gladkov Leonid A.
Candidate of Technical Sciences, Associate Professor, Southern Federal University, Taganrog, Russia
Gladkova Nadezhda V.
Senior Lecturer, Southern Federal University, Taganrog, Russia

Abstract

The article proposes a hybrid approach to solving optimization problems of computeraided design. The formulation of the problem, limitations and a criterion for assessing the quality of the solutions obtained are presented. The prospects for the integration of methods of computational intelligence and multiagent systems are shown. A hybrid approach is described based on a combination of evolutionary search methods, fuzzy logic, and the ability to parallelize the computational process. A modified migration operator for interpopulation information exchange has been developed. A scheme for organizing a parallel search process is proposed. The software implementation of the proposed algorithm has been completed. A brief description of the computational experiments carried out confirming the effectiveness of the proposed method is presented

Keywords

design tasks, bioinspired algorithms, neural networks, hybrid methods, parallel computing