Abstract—Intelligent Agriculture (IA) system is a complex system featured by complexity, uncertainty and large time-delay, and all its subsystems need cooperation with feedbacks from each other so that the whole system's control targets could be achieved. Components of IA system are dispersed unities with independent control targets, it is proved that Multi-Agent technology for system modeling and control is a useful method with simplicity and validity. In this paper, we had tried to research the cooperative control for IA system with this method firstly, and the way of Q-learning was used in researching of multi-agent collaboration control inference rule. According to control needs of three kinds of strawberries planted in the same greenhouse, we had designed different control agents corresponding to each environment variables. Joint optimal solution among these factors had been achieved through global control optimizing by cooperation controller, and intelligent adjusting of whole system can be effectively realized.

Index Terms—Intelligent agriculture (IA), complex system, reinforcement learning, cooperative control.

The authors are with the College of Information and Electrical Engineering, China Agricultural University, 100083, Beijing, China (e-mail:glhfei@126.com, tel.: + 86-10-62344012; fax: +86-10-62329064).

Cite: Ge Li, Yifei Chen, and Shangfeng Du, "Research on Cooperative Control in Intelligent Agriculture Complex System," International Journal of Computer and Communication Engineering vol. 1, no. 1, pp. 19-21 , 2012.