非线性最优化一个可行序列等式约束二次规划算法

A Feasible Equality Constrained SQP Algorithm for Nonlinear Optimization

本文针对非线性不等式约束优化问题,提出了一个新的可行序列等式约束二次规划算法．在每次迭代中,该算法只需求解三个相同规模且仅含等式约束的二次规划(必要时求解一个辅助的线性规划),因而其计算工作量较小．在一般的条件下,证明了算法具有全局收敛及超线性收敛性．数值实验表明算法是有效的．

In this paper, a feasible sequential equality constrained quadratic programming algorithm is proposed to solve the nonlinear inequality constrained optimization. Per single iteration, it is only neccessary to sovle three equality constrained quadratic programmings with the same scale （or in addition, a linear programming）, so the computational effort is reduced. The theoretical analysis shows that the algorithm is global and superlinear convergence under some suitable conditions. Numerical results show that the method in this paper is effective.