摘要
This paper presents a trust region algorithm for nonlinear problems of equalities and inequalities. The problem is changed to a least-squares problem with non-negative constraints by using slack variables. An equivalent KKT condition is derived, which is the base of constracting the new algorithm. The global conver- gence is established under general conditions, and a local quadratic convergence is proved for nondegenerate point satisfying second-order sufficient conditions. Numerical tests for our algorithm are also presented.
This paper presents a trust region algorithm for nonlinear problems of equalities and inequalities. The problem is changed to a least-squares problem with non-negative constraints by using slack variables. An equivalent KKT condition is derived, which is the base of constracting the new algorithm. The global conver- gence is established under general conditions, and a local quadratic convergence is proved for nondegenerate point satisfying second-order sufficient conditions. Numerical tests for our algorithm are also presented.
出处
《数值计算与计算机应用》
CSCD
北大核心
2001年第1期53-62,共10页
Journal on Numerical Methods and Computer Applications
关键词
非线性等式
非线性不等式
信赖域算法
数值计算
trust region, nonlinear problems of equalities and inequalities, least square problem, quadratic convergence
