• Electronics Technology •

Multi-agent system application in accordance with game theory in bi-directional coordination network model

Jie ZHANG1,*(), Gang WANG2(), Shaohua YUE2(), Yafei SONG2(), Jiayi LIU2(), Xiaoqiang YAO2()

1. 1 College of Electronics and Information Engineering, Air Force Engineering University, Xi'an 710054, China
2 College of Air Missile Defense, Air Force Engineering University, Xi'an 710054, China
• Received:2019-04-30 Online:2020-04-30 Published:2020-04-30
• Contact: Jie ZHANG E-mail:afeu_zhangjie@163.com;sharesunny123@163.com;zhouguoan@sina.cn;yafei_song@163.com;sixandone1@163.com;icemissile@sina.com
• About author:ZHANG Jie was born in 1995. He is a master degree candidate at the Air Force Engineering University. His research interests are combat multi-agent based on deep learning and tactical air defense and antimissile command and control system.E-mail: afeu_zhangjie@163.com|WANG Gang was born in 1975. He received his Ph.D. degree from the Air Force Engineering University. His research interests are machine learning, information fusion and command and control system. E-mail: sharesunny123@163.com|YUE Shaohua was born in 1968. She received her Ph.D. degree from the Air Force Engineering University. Her research interests are command information system and intelligent command and control. E-mail: zhouguoan@sina.cn|SONG Yafei was born in 1988. He received his Ph.D. degree from the Air Force Engineering University. His research interests are pattern recognition and intelligent information processing. E-mail: yafei_song@163.com|LIU Jiayi was born in 1996. He is a master degree candidate at the Air Force Engineering University. His research interests are air defense and anti-missile command and control system and intelligent decision-making based on reinforcement learning. E-mail: sixandone1@163.com|YAO Xiaoqiang was born in 1985. He received his Ph.D. degree from the Air Force Engineering University. His research interests are intelligent information processing and simulation training and simulation.E-mail: icemissile@sina.com
• Supported by:
the National Natural Science Foundation of China(61503407);the National Natural Science Foundation of China(61806219);the National Natural Science Foundation of China(61703426);the National Natural Science Foundation of China(61876189);the National Natural Science Foundation of China(61703412);the China Postdoctoral Science Foundation(2016 M602996);This work was supported by the National Natural Science Foundation of China (61503407; 61806219; 61703426; 61876189; 61703412) and the China Postdoctoral Science Foundation (2016 M602996)

Abstract:

The multi-agent system is the optimal solution to complex intelligent problems. In accordance with the game theory, the concept of loyalty is introduced to analyze the relationship between agents' individual income and global benefits and build the logical architecture of the multi-agent system. Besides, to verify the feasibility of the method, the cyclic neural network is optimized, the bi-directional coordination network is built as the training network for deep learning, and specific training scenes are simulated as the training background. After a certain number of training iterations, the model can learn simple strategies autonomously. Also, as the training time increases, the complexity of learning strategies rises gradually. Strategies such as obstacle avoidance, firepower distribution and collaborative cover are adopted to demonstrate the achievability of the model. The model is verified to be realizable by the examples of obstacle avoidance, fire distribution and cooperative cover. Under the same resource background, the model exhibits better convergence than other deep learning training networks, and it is not easy to fall into the local endless loop. Furthermore, the ability of the learning strategy is stronger than that of the training model based on rules, which is of great practical values.