A nonmonotone line search technique and its application to unconstrained optimization

被引:524
作者
Zhang, HC [1 ]
Hager, WW [1 ]
机构
[1] Univ Florida, Dept Math, Gainesville, FL 32611 USA
关键词
nonmonotone line search; R-linear convergence; unconstrained optimization; L-BFGS method;
D O I
10.1137/S1052623403428208
中图分类号
O29 [应用数学];
学科分类号
070104 ;
摘要
A new nonmonotone line search algorithm is proposed and analyzed. In our scheme, we require that an average of the successive function values decreases, while the traditional nonmonotone approach of Grippo, Lampariello, and Lucidi [SIAM J. Numer. Anal., 23 ( 1986), pp. 707 - 716] requires that a maximum of recent function values decreases. We prove global convergence for nonconvex, smooth functions, and R-linear convergence for strongly convex functions. For the L-BFGS method and the unconstrained optimization problems in the CUTE library, the new nonmonotone line search algorithm used fewer function and gradient evaluations, on average, than either the monotone or the traditional nonmonotone scheme.
引用
收藏
页码:1043 / 1056
页数:14
相关论文
共 17 条
[1]   2-POINT STEP SIZE GRADIENT METHODS [J].
BARZILAI, J ;
BORWEIN, JM .
IMA JOURNAL OF NUMERICAL ANALYSIS, 1988, 8 (01) :141-148
[2]   Nonmonotone spectral projected gradient methods on convex sets [J].
Birgin, EG ;
Martínez, JM ;
Raydan, M .
SIAM JOURNAL ON OPTIMIZATION, 2000, 10 (04) :1196-1211
[3]   CUTE - CONSTRAINED AND UNCONSTRAINED TESTING ENVIRONMENT [J].
BONGARTZ, I ;
CONN, AR ;
GOULD, N ;
TOINT, PL .
ACM TRANSACTIONS ON MATHEMATICAL SOFTWARE, 1995, 21 (01) :123-160
[4]   R-linear convergence of the Barzilai and Borwein gradient method [J].
Dai, YH ;
Liao, LZ .
IMA JOURNAL OF NUMERICAL ANALYSIS, 2002, 22 (01) :1-10
[5]  
Dai Yuhong, 2002, Journal of Systems Science and Complexity, V15, P139
[6]   A TRUNCATED NEWTON METHOD WITH NONMONOTONE LINE SEARCH FOR UNCONSTRAINED OPTIMIZATION [J].
GRIPPO, L ;
LAMPARIELLO, F ;
LUCIDI, S .
JOURNAL OF OPTIMIZATION THEORY AND APPLICATIONS, 1989, 60 (03) :401-419
[7]   A CLASS OF NONMONOTONE STABILIZATION METHODS IN UNCONSTRAINED OPTIMIZATION [J].
GRIPPO, L ;
LAMPARIELLO, F ;
LUCIDI, S .
NUMERISCHE MATHEMATIK, 1991, 59 (08) :779-805
[8]   A NONMONOTONE LINE SEARCH TECHNIQUE FOR NEWTON METHOD [J].
GRIPPO, L ;
LAMPARIELLO, F ;
LUCIDI, S .
SIAM JOURNAL ON NUMERICAL ANALYSIS, 1986, 23 (04) :707-716
[9]   ON THE LIMITED MEMORY BFGS METHOD FOR LARGE-SCALE OPTIMIZATION [J].
LIU, DC ;
NOCEDAL, J .
MATHEMATICAL PROGRAMMING, 1989, 45 (03) :503-528
[10]   Curvilinear stabilization techniques for truncated Newton methods in large scale unconstrained optimization [J].
Lucidi, S ;
Rochetich, F ;
Roma, M .
SIAM JOURNAL ON OPTIMIZATION, 1998, 8 (04) :916-939