Abstract: Aiming at the conservative problem of uncertain discrete-time output Probability Density Fuctions（PDFs）control systems, guaranteed cost control algorithm based on Linear Matrix Inequality(LMI)neural control was proposed, to which added state feedback controller of additive gain. In controller design, neural controller was used to tune the additive gain so that the guaranteed cost was reduced, and feasibility conditions for guaranteed cost control algorithm were derived by using LMI. The examples were used to demonstrate that the proposed algorithm could reduce the conservation of system and improve the stability and robustness of system.

Key words: probability density fuctions, guaranteed cost, linear matrix inquality, neurocontroller, additive gain

摘要： 针对离散不确定输出概率密度函数控制系统存在的保守性问题，提出基于线性矩阵不等式神经控制的保性能控制算法。该控制算法是在线性矩阵不等式设计的基础上，加入附加增益的状态反馈控制，在控制器设计时用神经控制器调控附加增益，降低系统的保守性，同时利用线性矩阵不等式的方法给出保性能控制算法的可行性条件。通过实验证明，该保性能控制算法能够降低系统保守性，能较好地改善系统鲁棒性和稳定性。

关键词: 概率密度函数, 保性能, 线性矩阵不等式, 神经控制器