ON THE RESOLUTION OF LINEARLY CONSTRAINED CONVEX MINIMIZATION PROBLEMS

被引:20
作者
FRIEDLANDER, A
MARTINEZ, JM
SANTOS, SA
机构
关键词
LARGE-SCALE LINEARLY CONSTRAINED OPTIMIZATION; BOX-CONSTRAINED PROBLEMS; OPTIMALITY CONDITIONS;
D O I
10.1137/0804018
中图分类号
O29 [应用数学];
学科分类号
070104 ;
摘要
The problem of minimizing a twice differentiable convex function f is considered, subject to Ax = b, x greater than or equal to 0, where A is an element of IR(MxN), M, N are large and the feasible region is bounded. It is proven that this problem is equivalent to a ''primal-dual'' box-constrained problem With 2N + M variables. The equivalent problem involves neither penalization parameters nor ad hoc multiplier estimators. This problem is solved using an algorithm for bound constrained minimization that can deal with many variables. Numerical experiments are presented.
引用
收藏
页码:331 / 339
页数:9
相关论文
共 17 条
[1]  
Conn A.R., 1991, 9110 FAC U NAM DEP M
[2]   A GLOBALLY CONVERGENT AUGMENTED LAGRANGIAN ALGORITHM FOR OPTIMIZATION WITH GENERAL CONSTRAINTS AND SIMPLE BOUNDS [J].
CONN, AR ;
GOULD, NIM ;
TOINT, PL .
SIAM JOURNAL ON NUMERICAL ANALYSIS, 1991, 28 (02) :545-572
[3]   CORRECTION [J].
CONN, AR .
SIAM JOURNAL ON NUMERICAL ANALYSIS, 1989, 26 (03) :764-767
[4]  
CONN AR, 1988, MATH COMPUT, V50, P399, DOI 10.1090/S0025-5718-1988-0929544-3
[5]   GLOBAL CONVERGENCE OF A CLASS OF TRUST REGION ALGORITHMS FOR OPTIMIZATION WITH SIMPLE BOUNDS [J].
CONN, AR ;
GOULD, NIM ;
TOINT, PL .
SIAM JOURNAL ON NUMERICAL ANALYSIS, 1988, 25 (02) :433-460
[6]  
Duff I. S., 2017, DIRECT METHODS SPARS
[7]  
Fletcher R., 1981, PRACTICAL METHODS OP
[8]  
FRIEDLANDER A, 1992, NEW TRUST REGION ALG
[9]  
FRIEDLANDER A, IN PRESS SIAM J OPTI
[10]  
GREEN PJ, 1990, J ROY STAT SOC B MET, V52, P443