论文部分内容阅读
本文通过线性最优控制理论、灵敏度分析理论及优化理论的有机结合,提出了一种系统的权矩阵优化选取方法.本方法在设计人员给定主导特征根位置、反馈增益顶值限制、寻优步长等条件下自动寻优,求得满足给定条件的权矩阵.同时为了最大限度地发挥权矩阵的作用,采用对权矩阵进行满阵优化的方法,大大提高了优化效率。实例表明,本方法能方便而快速地求得所需的权矩阵,解除了设计人员通过试凑法选取合适权矩阵的沉重负担.特别是用本方法所设计的降阶反馈控制系统较之一般常用的最小误差激励法所得到的降阶反馈控制系统具有更优良的动态品质.
In this paper, an optimization method of weight matrix is proposed based on the combination of linear optimal control theory, sensitivity analysis theory and optimization theory.The method is based on the fact that the designer gives the dominant root location, feedback gain top limit, Step size and other conditions to find the weight matrix which satisfies the given conditions.At the same time, in order to maximize the function of weight matrix, the method of full matrix optimization of weight matrix is adopted, which greatly improves the optimization efficiency. The example shows that this method can get the weight matrix conveniently and quickly, and relieves the designers from the heavy burden of choosing the right matrix through the trial and error method.Especially the degenerated feedback control system designed by this method is better than the general one The commonly used method of minimum error excitation feedback control system has a better dynamic quality.