针对非线性模型预测控制中滚动优化问题在线求解的困难,提出一种基于可行扰动的实时迭代优化算法。将可行扰动策略引入滚动优化中,保证算法快速收敛且具有提前终止能力。通过对运动小车以及连续搅拌反应釜的仿真研究,验证了该算法的有效性。
A real time iterative algorithm based on feasible perturbation was developed to deal with the difficulties for on-line receding-horizon optimization problems in nonlinear model predictive control.The feasible perturbation was introduced into receding-horizon optimization problems to guarantee fast convergence and early termination.Simulation results of moving cart and continuous stirred tank reactor showed the effectiveness of the proposed algorithm.