综合简约Hessian SQP算法及在多元精馏优化中的应用

针对化工系统流程优化问题中决策变量子空间维数低的特点，提出了基于坐标基分解的综合简约Ｈｅｓｓｉａｎ ＳＱＰ算法，并直接用于大规模优化问题求解；同时，在算法中引入拟拉格朗日函数的概念，用有限差分法综合处理简约ＱＰ问题目标函数的一次项系数，进一步提高了优化计算和求解的可靠性。

INTEGRATED REDUCED HESSIAN SQP ALGORITHM FOR DISTILLATION COLUMN OPTIMIZATION

To extend successive quadratic programming(SQP) to handle large scale chemical process optimization successfully,a set of integrated reduced Hessian SQP (IRHSQP) algorithm based on coordinate bases decomposition is developed,which exploits the low dimensionality of the subspace of the decision variables.With great efficiency and reliability,the resulted algorithm presents the definition of quasi Lagrange function and simultaneously deals with the first order term of reduced QP problem with finite difference approach.Finally,a numerical comparison of algorithm with traditional strategies is shown on a distillation column optimization problem.