迭代学习型瞬时最优控制及其收敛性分析

将传统的瞬时最优化控制和智能算法中的迭代学习控制相结合,提出了基于最优化控制算法和智能控制算法的迭代学习型瞬时最优化控制算法.该方法以线性系统为模型,以系统的响应与期望响应的差值为反馈,以二次型性能泛函为目标函数,通过迭代学习修正主动控制器的控制信号,提高主动控制的效果.针对迭代学习型瞬时最优化控制算法迭代的特性,本文采用范数方法给出了该方法收敛的充分条件.为验证方法的有效性,选取第二代基准模型作为计算模型,埃尔森特罗地震波南北分量作为输入载荷,数值仿真结果表明,迭代学习型瞬时最优控制算法较传统的瞬时最优控制算法有更好的控制效果.

OPTIMAL CONTROL USING INSTANTANEOUS OPTIMAL AND ITERATIVE LEARNING CONTROL AND THE CONVERGENCE ANALYSIS

By combining instantaneous optimal control and iterative learning control(ILC), one new hybrid control strategy called instantaneous optimal iterative learning control is proposed. Linear system is chosen as the model for the new control strategy, and the quadratic performance function of the system is chosen as the objective function to be minimized. During the process of controlling responses of the system, the core idea of the iterative learning control is introduced in order to modify the control signals. By introducing the norms of matrices, the sufficient condition of convergence for the new control strategy is established in the paper. The model of a 20-floor building in the second generation benchmark vibration control is selected for numerical simulation. In the numerical simulation, the north-south component of the El wave is introduced as the excitation. Comparing to the instantaneous optimal control, results of the simulation show that instantaneous optimal iterative learning control improves the effectiveness.