基于MAS的分布式焦炉集气管压力解耦控制

针对焦炉集气管压力这类多变量非线性耦合系统,提出了一种基于multi-agent system(MAS)的焦炉集气管压力智能协调控制系统方案,给出了agent的分层组织结构和演化机制．在控制agent中,采用TS模糊神经回归网络对被控对象进行分布式建模,运用分布式智能协调解耦算法进行解耦控制,监督学习与强化学习相结合,采用遗传协进化算法对多个agent协调优化．通过agent模态变迁进行模式切换,以适应快速突变环境．工程应用表明,提出的控制策略有效地解决了集气管压力这类复杂对象的过程控制问题．

MAS-based distributed decoupling control for the pressure of gas collectors of coke ovens

For the multi-variable nonlinear coupled system such as the pressure of gas collectors of coke ovens,a scheme of intelligent coordinated control system is proposed based on multi-agent system.Firstly,the structures of hierarchical organization and evolution mechanism of agents in system are presented.In control agent,the T-S (Takagi-Sugeno) type recurrent fuzzy network (TSRFN) is employed to set up the distributed model of the plant,and the distributed intelligent coordinated decoupling control algorithm is adopted to realize the decoupling control.The supervised learning combines with reinforcement learning.The agents in system are then optimized coordinately by using the genetic co-evolution algorithm.The system can switch to different mode using the state-change of agents in order to operate in rapid time- varying environments.Finally,the real-world application shows that the proposed control strategy successfully solved the complex process control problem such as the pressure of gas collectors of coke ovens.