Control of technical object on the basis of the multi-agent system with neuroevolution and student-teacher of-line learning

7th International Scientific and Research Conference "Topical Issues in Aeronautics and Astronautics"

航空航天工程

Engel, E.A.^1 ; Kovalev, I.V.^2 ; Engel, N.E.^1

First Russian Doctoral Degree in Computer Sciences, Katanov State University of Khakassia, Abakan, Russia^1

Second Russian Doctoral Degree in Computer Sciences, Siberian State Aerospace University, Krasnoyarsk, Russia^2

关键词: High quality;    Learning and teachings;    Near-optimal control;    Neuro evolutions;    Simulation studies;    Student teachers;    Technical objects;    Traditional learning;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/155/1/012001/pdf
DOI  :  10.1088/1757-899X/155/1/012001

学科分类：航空航天科学

来源: IOP

【 摘 要 】

This paper presents algorithm for generating neuroevolutionary multi-agent system that allows agents to learn from high-quality activities. Dissimilar traditional learning algorithms proposed algorithm combines student-teacher of-line learning and teaching agents based on sufficient activities producing by any agent in its subculture. The simulation studies demonstrated that the proposed algorithm is effective at rapidly generating near-optimal control agents.