基于改进相对熵的控制系统性能评价方法

Performance Assessment Method of Control Systems Based on Improved Relative Entropy

为提高控制系统随机性能评价的准确性,克服传统相对熵指标非对称、无界的问题,提出一种基于改进相对熵指标(IREI)的控制系统性能评价方法。该方法首先将原始相对熵指标改进为对称有界的测度指标,即IREI;然后采用核密度估计法计算过程数据的概率密度函数,通过引入蒙特卡罗模拟以逼近非高斯分布情况下的指标值。最后,利用MATLAB/Simulink建立包含非高斯噪声扰动及调节阀劣化的仿真模型,设计一致性检验与随机扰动检验实验,将IREI与确定性指标及最小方差指标进行对比。对比结果表明,该方法可以灵活有效地评价控制器性能。

In order to improve the accuracy of random performance assessment of control systems and overcome the problem of asymmetry and boundlessness of traditional relative entropy index,a performance assessment method of control systems based on improved relative entropy index(IREI)is proposed.Firstly,the original relative entropy index is improved to be a symmetric and bounded measure index,namely IREI.Then,the kernel density estimation is used to obtain the probability density function of the process data.Monte Carlo simulation is introduced to approximate the index value in the case of non-Gaussian distribution.Finally,MATLAB/Simulink was used to establish a simulation model including non-Gaussian noise disturbance and control valve deterioration.The experiments of consistency test and random disturbance test are conducted to compare the IREI with the deterministic index and minimum variance index.The comparison results show that the IREI can assess the performance of the controller effectively and flexibly.