鲁棒辨识的l_1和H_∞误差上界的估计算法

Computing l_1 and H_∞ error bound in robust identification with intelligent algorithm

对有限参数线性系统辨识问题的l1误差上界估计和时域H∞ 插值算法误差上界估计等问题转化为一类分片线性函数的最优化问题 ,提出基于分片的混合遗传算法 .吸取进化算法的探索能力 ,并充分利用模型信息 .仿真实验结果表明了它的寻优性能 .

The problem of estimating both the \$l\-1\$ upper error bound for robust identification and upper error bound for H\-∞ interpolation algorithms is formulated into the optimization of piece-wise linear functions subjected to linear constrains. Propose a piece-wise based hybrid genetic algorithm. Its performance is proved by simulation experiments.