A trust region algorithm for optimization with nonlinear equality and linear inequality constraints

被引：0

作者：

陈中文

韩继业

机构：

[1] China

[2] Chinese Academy of Sciences

[3] Institute of Applied Mathematics

[4] Beijing 100080

来源：

Science China Mathematics | 1996年 / 08期

基金：

中国国家自然科学基金;

关键词：

trust regions; global convergence; nonlinear equality constraints; linear inequality constraints;

D O I：

暂无

中图分类号：

O224 [最优化的数学理论];

学科分类号：

070105 ; 1201 ;

摘要：

A new algorithm of trust region type is presented to minimize a differentiable function ofmany variables with nonlinear equality and linear inequality constraints. Under the milder conditions, theglobal convergence of the main algorithm is proved. Moreover, since any nonlinear inequality constraint can beconverted into an equation by introducing a slack variable, the trust region method can be used in solving general nonlinear programming problems.

引用

页码：799 / 806

页数：8

共 4 条

[1] A NEW TRUST REGION ALGORITHM FOR BOUND CONSTRAINED MINIMIZATION [J].

FRIEDLANDER, A ;

MARTINEZ, JM ;

SANTOS, SA .

APPLIED MATHEMATICS AND OPTIMIZATION, 1994, 30 (03) :235-266

[2]

A trust region algorithm for equality constrained optimization[J] . M. J. D. Powell,Y. Yuan.Mathematical Programming . 1990 (1)

[3]

Convergence properties of trust region methods for linear and convex constraints[J] . James V. Burke,Jorge J. Moré,Gerardo Toraldo.Mathematical Programming . 1990 (1)

[4] ON THE GLOBAL CONVERGENCE OF TRUST REGION ALGORITHMS FOR UNCONSTRAINED MINIMIZATION [J].

POWELL, MJD .

MATHEMATICAL PROGRAMMING, 1984, 29 (03) :297-303

← 1 →