基于多项式核的稀疏核学习单步预测控制算法及其应用

One-step-ahead predictive control based on sparse kernel learning with polynomial kernel and its application to chemical processes

提出一种基于稀疏核学习辨识模型的单步预测控制(sparse kernel learningone-step-ahead predictive control,SKL-OPC)框架,并推导了该框架下采用多项式核的一种控制算法。该算法在求取最优控制律时可将调节变量从目标函数分离出来,并最终转化为求解一奇数次代数方程根的问题。因此无需复杂的非线性优化技术,且克服了基于二次多项式核辨识模型不准确造成控制算法失效的缺点。在一非线性连续搅拌反应釜的控制研究表明了该方法的有效性和优越性。

A novel control framework based on sparse kernel learning one-step ahead predictive control （SKL-OPC） was presented for the general unknown nonlinear processes. The polynomial kernel function was adopted to derive a simple control strategy as a natural extension of SKL-OPC. The manipulated input can be separated from the control performance index due to the special structure of polynomial kernel. Consequently, the control problem was resolved by solving the roots of an odd degree polynomial equation. The proposed control strategy did not require the nonlinear optimization technique, which resulted in a small computation scale and made it very suitable for real-time control. Application of the proposed approach to a highly nonlinear continuous stirred tank reactor indicated its validity and showed superior performance, compared to other methods.