摘要
设计了求解不等式约束非线性规划问题的一种新的滤子序列线性方程组算法.该算法每步迭代由减小约束违反度和目标函数值两部分构成.利用约束函数在某个中介点线性化的方法产生搜索方向.每步迭代仅需求解两个线性方程组,计算量较小.在一般条件下,证明了算法产生的无穷迭代点列所有聚点都是可行点并且所有聚点都是所求解问题的KKT点.
A new globally convergent filter sequential systems of linear equation algorithm is presented for inequality constrained optimization problems. Each iteration of this algorithm is composed of a feasibility phase, which reduces a measure of infeasibility, and an optimality phase,which reduces the objective function. A search direction is obtained by linearization of the feasible set at an intermediate point. In each iteration we only need to solve two linear equations. Hence fewer computations are required. Under mild assumptions it is showed that every limit point of the sequence of iterates in this algorithm is feasible,and also is a KKT point.
出处
《应用数学》
CSCD
北大核心
2010年第3期602-609,共8页
Mathematica Applicata
基金
江苏省农机局科研基金(GXZ06014)
