大规模过程系统优化的简约空间SQP算法

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．

Reduced-Space SQP Algorithm Based on Orthogonal Decomposition for Large-Scale Process Optimization

Reduced space SQP algorithm for the optimization of large scale process systems are studied in this paper. Variables decomposition is used to eliminate dependent variables and equality constraints, which reduce the dimension of QP sub problems involved in iterations of SQP. Compared with those based on orthonormal bases, reduced space SQP based on orthogonal bases retains the convergence property while avoids QR decomposition. The improved computation efficiency is illustrated in the distillation column optimization problem.