The efficiency of subgradient projection methods for convex optimization .1. General level methods

被引:46
作者
Kiwiel, KC
机构
[1] Systems Research Institute, 01-447 Warsaw
关键词
nondifferentiable (nonsmooth) optimization; convex programming; relaxation methods; subgradient optimization; successive projections; linear inequalities; parallel computing;
D O I
10.1137/0334031
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
We study subgradient methods for convex optimization that use projections onto successive approximations of level sets of the objective corresponding to estimates of the optimal value. We present several variants and show that they enjoy almost optimal efficiency estimates. In another paper we discuss possible implementations of such methods. In particular, their projection subproblems may he solved inexactly via relaxation methods, thus opening the way for parallel implementations. They can also exploit accelerations of relaxation methods based on simultaneous projections, surrogate constraints, and conjugate and projected (conditional) subgradient techniques.
引用
收藏
页码:660 / 676
页数:17
相关论文
共 19 条
[11]  
KULIKOV AN, 1990, ZH VYCHISL MAT MAT F, V30, P663
[12]   NEW VARIANTS OF BUNDLE METHODS [J].
LEMARECHAL, C ;
NEMIROVSKII, A ;
NESTEROV, Y .
MATHEMATICAL PROGRAMMING, 1995, 69 (01) :111-147
[13]  
LEMARECHAL C, 1991, 1508 INRIA
[14]  
Lemarechal C., 1989, Handbooks in Operations Research and Management Science, P529, DOI [DOI 10.1016/S0927-0507(89)01008-X, 10.1016/s0927-0507(89)01008-x]
[15]   POLYNOMIAL ALGORITHMS FOR A CLASS OF LINEAR-PROGRAMS [J].
MAURRAS, JF ;
TRUEMPER, K ;
AKGUL, M .
MATHEMATICAL PROGRAMMING, 1981, 21 (02) :121-136
[16]  
MOTZKIN TS, 1954, CAN J MATH, V6, P393, DOI 10.4153/CJM-1954-038-x
[17]  
Nemirovskii Arkadii, 1983, A Wiley-Interscience publication
[18]  
Polyak BT., 1969, USSR COMP MATH MATH, V9, P14, DOI DOI 10.1016/0041-5553(69)90061-5
[19]   ON RELAXATION METHODS FOR SYSTEMS OF LINEAR INEQUALITIES [J].
TELGEN, J .
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH, 1982, 9 (02) :184-189