不等式约束优化一个具有超线性收敛的可行序列二次规划算法

A FEASIBLE SQP ALGORITHM WITH SUPERLINEAR CONVERGENCE FOR INEQUALITY CONSTRAINED OPTIMIZATION

建立了一个新的SQP算法,提出了一阶可行条件这一新概念．对已有SQP型算法进行改进,减少计算工作量,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, we present a new SQP algorithm, and a new idea called first-order feasible condition, by which the traditional SQP type method is improved, and computational effort is reduced. Theoretical analysis shows that the algorithm is global and superlinear convergent under some suitable conditions. The numerical results show that the method in this paper is effective.