论文部分内容阅读
提出了一种适用于大规模过程系统优化的简约空间ＳＱＰ方法．简约空间ＳＱＰ算法用变量分解的方法消去非独立变量和等式约束，以降低ＱＰ子问题的维数，使每次迭代的ＱＰ子问题求解只在零空间中进行．正交基分解与标准正交基分解相比，具有同样的收敛性能，但由于避免了ＱＲ分解，故进一步减少了计算量，是大规模过程系统优化中很有前途的算法．优化算例说明了这一方法的有效性．
A simple space SQP method suitable for large-scale process system optimization is proposed. The simple space SQP algorithm eliminates the independent variables and equality constraints by variable decomposition to reduce the dimension of QP sub-problems so that the QP sub-problems in each iteration are solved in null space only. The orthogonal basis decomposition has the same convergence performance compared with the standard orthogonal basis decomposition, but further simplifies the system optimization because of avoiding the QR decomposition. The optimization example illustrates the effectiveness of this method.