摘要
针对火电厂非线性、多变量和多控制目标的特点,设计了一个火电厂多代理控制系统(PPMACS).在PP MACS中,前馈控制代理(FFCAs)采用神经模糊系统进行决策,反馈控制代理(FBCAs)采用基于遗传算法的模糊系统进行决策.优化任务分解代理(OTDAs)通过一个优化代理和一个分解代理来进行多目标优化分解PPMACS的任务.协调代理根据运行条件协调PPMACS的各个代理.仿真结果显示了火电厂多代理控制系统能够实现火电单元机组的多目标运行和大范围负荷跟踪.神经网络、模糊逻辑和遗传算法是PPMACS中的智能代理进行决策的有效工具.
To deal with the nonlinear multi-variable and multiple control objective characteristic of power plant, the power plant multiagent control system (PPMACS) is designed. In the PPMACS, the feedforward control agents (FFCAs) make decisions using the neuro-fuzzy systems and the feedback control agents (FBCAs) make decisions using the genetic algorithm-based fuzzy \{systems\}. The optimal task decomposition agents (OTDAs) optimally decompose the task of the PPMACS through an optimization agent and a decomposition agent. The coordinator agent (COA) coordinates the agents in the PPMACS according to different \{operating\} conditions. Simulation results demonstrate that the PPMACS implement the multiobjective operation and wide range load tracking.Neural netwroks,fuzzy logic and genetic algorithm are effective tools for the agents of the PPMACS in making decisions.
出处
《控制理论与应用》
EI
CAS
CSCD
北大核心
2004年第3期405-410,共6页
Control Theory & Applications
关键词
决策
反馈控制
前馈控制
模糊逻辑
遗传算法
多代理控制系统
神经网络
decision-making
feedback control
feedforward control
fuzzy logic
genetic algorithms
multiagent control system
neural networks