NUMERICAL EXPERIENCE WITH A CLASS OF ALGORITHMS FOR NONLINEAR OPTIMIZATION USING INEXACT FUNCTION AND GRADIENT INFORMATION

被引:27
作者
CARTER, RG [1 ]
机构
[1] NASA,LANGLEY RES CTR,INST COMP APPLICAT SCI & ENGN,HAMPTON,VA 23665
关键词
UNCONSTRAINED OPTIMIZATION; TRUST REGION METHODS; INEXACT FUNCTION EVALUATION; INEXACT GRADIENTS; VARIABLE-ACCURACY SIMULATIONS;
D O I
10.1137/0914023
中图分类号
O29 [应用数学];
学科分类号
070104 ;
摘要
For optimization problems associated with engineering design, parameter estimation, image reconstruction, and other optimization/simulation applications, low accuracy function and gradient values are frequently much less expensive to obtain than high accuracy values. The computational performance of trust region methods for nonlinear optimization is investigated for cases when high accuracy evaluations are unavailable or prohibitively expensive, and earlier theoretical predictions that such methods are convergent even with relative gradient errors of 0.5 or more is confirmed. The proper choice of the amount of accuracy to use in function and gradient evaluations can result in orders-of-magnitude savings in computational cost.
引用
收藏
页码:368 / 388
页数:21
相关论文
共 19 条
[1]  
[Anonymous], 1970, NONLINEAR PROGRAMMIN, DOI DOI 10.1016/B978-0-12-597050-1.50006-3
[2]  
Apostol T.M., 1957, MATH ANAL
[3]   ON THE GLOBAL CONVERGENCE OF TRUST REGION ALGORITHMS USING INEXACT GRADIENT INFORMATION [J].
CARTER, RG .
SIAM JOURNAL ON NUMERICAL ANALYSIS, 1991, 28 (01) :251-265
[4]  
CARTER RG, 1989, NASA8945 I COMP APPL
[5]  
CARTER RG, 1987, TR8706 RIC U DEP MAT
[6]  
DENNIS JE, 1983, NUMERICAL METHODS UN
[7]  
FLETCHER R, 1982, NONLINEAR OPTIMIZATI, P185
[8]   COMPUTING OPTIMAL LOCALLY CONSTRAINED STEPS [J].
GAY, DM .
SIAM JOURNAL ON SCIENTIFIC AND STATISTICAL COMPUTING, 1981, 2 (02) :186-197
[9]  
Hindmarsh A.C., 1983, SCI COMPUT, P55
[10]  
Lambert J.D, 1973, COMPUTATIONAL METHOD