非线性互补约束均衡问题的改进SQP算法

摘要

In this thesis, we make an investigation on feasible sequential quadratic programming algorithm. At first, complementary constraints are smoothed with the idea of successive approximation,and the mathematical program ming with equilibrium constraints is changed equivalently a standard smooth nonlinear programming. Then, the smooth nonlinear programming is solved by the idea of above-mentioned sequential quadratic programming algo rithms. The existence and detailed solution procedure of the auxiliary direction for the SQP method is presented, which makes the existing algorithm more reasonable and feasible.%为了对可行序列二次规划算法进行研究与创新,利用逐步逼近思想,对互补约束条件光滑化,将均衡问题等价转化为一个光滑的标准非线性规划问题,进而利用序列二次规划算法思想求解.给出了SQP算法中辅助方向的存在性分析和具体求解方法,使算法更加合理可行.