将克隆免疫算法与预测控制结合,提出了一种免疫预测控制方法.该算法利用克隆选择算法实现预测控制的滚动优化,避免了求Diophantine方程及逆矩阵的复杂推导.仿真结果表明,该方法对时滞系统、非最小相位系统、不稳定对象、非线性系统、MIMO系统都能达到理想的控制效果,对外部干扰及建模误差具有很好的适应性.
An immune predictive control method based on Clone Selection Algorithm has been proposed. The algorithm is based on the Auto Regressive Moving Average (ARMA) model and rolling horizon optimization by using clone selection, and can avoid the complexities in solving the Diophantine equations. Simulation results have shown that the algorithm can effectively handle the time delay system, non-minimum phase of the process, the instability process, MIMO and nonlinear systems, and can effectively cope with disturbance and model uncertainty.