智能控制：从学习控制到平行控制
Intelligent control: from learning control to parallel control

20世纪60年代, 学习控制开启了人类探究复杂系统控制的新途径, 基于人工智能技术的智能控制随之兴起. 本 文以智能控制为主线, 阐述其由学习控制向平行控制发展的历程. 本文首先介绍学习控制的基本思想, 描述了智能机器 的架构设计与运行机理. 随着信息科技的进步, 基于数据的计算智能方法随之出现. 对此, 本文进一步简述了基于计算 智能的学习控制方法, 并以自适应动态规划方法为切入点分析非线性动态系统自学习优化问题的求解过程. 最后, 针对 工程复杂性与社会复杂性互相耦合的复杂系统控制问题, 阐述了基于平行控制的学习与优化方法求解思路, 分析其在求 解复杂系统优化控制问题方面的优势. 智能控制思想经历了学习控制、计算智能控制到平行控制的演化过程, 可以看出 平行控制是实现复杂系统知识自动化的有效方法.
In the 1960s, learning control opened up new ways for humans to explore the control of complex systems, and intelligent control emerged based on artificial intelligence technologies thereafter. In this paper, intelligent control is employed as the main line, which describes the development process of intelligent control from learning control to parallel control. First, the basic idea of learning control is introduced, and then the architecture design and operating mechanism of intelligent machines are described. With the advancement of information science, data-based computational intelligence approaches are emerged. Then, the learning control methods based on computational intelligence are described. Adaptive dynamic programming, as one of the most well-known self-learning optimal control methods, is described to show how to solve the optimal control problems for nonlinear dynamic systems. Finally, considering the control problem of complex systems with both engineering and social complexities, parallel control is introduced and its advantages in solving complex optimization control problems are illustrated. Intelligent control is developed from learning control to computational intelligence control and then to parallel control. It can be seen that parallel control is an effective method to realize the knowledge automation of complex systems.