UNIVERSAL REGULARIZATION METHODS: VARYING THE POWER, THE SMOOTHNESS AND THE ACCURACY

被引:37
作者
Cartis, Coralia [1 ]
Gould, Nick, I [2 ]
Toint, Philippe L. [3 ]
机构
[1] Univ Oxford, Math Inst, Oxford OX2 6GG, England
[2] STFC Rutherford Appleton Lab, Computat Sci & Engn Dept, Chilton OX11 0QX, England
[3] Univ Namur, Namur Res Ctr Complex Syst NAXYS, B-5000 Namur, Belgium
关键词
evaluation complexity; worst-case analysis; regularization methods; CASE EVALUATION COMPLEXITY; NEWTON METHODS; DESCENT;
D O I
10.1137/16M1106316
中图分类号
O29 [应用数学];
学科分类号
070104 ;
摘要
Adaptive cubic regularization methods have emerged as a credible alternative to linesearch and trust-region for smooth nonconvex optimization, with optimal complexity amongst second-order methods. Here we consider a general/new class of adaptive regularization methods that use first- or higher-order local Taylor models of the objective regularized by a(ny) power of the step size and applied to convexly constrained optimization problems. We investigate the worst-case evaluation complexity/global rate of convergence of these algorithms, when the level of sufficient smoothness of the objective may be unknown or may even be absent. We find that the methods accurately reflect in their complexity the degree of smoothness of the objective and satisfy increasingly better bounds with improving model accuracy. The bounds vary continuously and robustly with respect to the regularization power and accuracy of the model and the degree of smoothness of the objective.
引用
收藏
页码:595 / 615
页数:21
相关论文
共 26 条
  • [1] [Anonymous], 2000, Society for Industrial and Applied Mathematics
  • [2] [Anonymous], 1983, WILEY SER DISCRETE M
  • [3] [Anonymous], 2017, PREPRINT
  • [4] [Anonymous], 2017, ARXIV171011606
  • [5] Bensoussan A., 2002, REGULARITY RESULTS N
  • [6] Bertsekas D., 1999, NONLINEAR PROGRAMMIN
  • [7] On the use of third-order models with fourth-order regularization for unconstrained optimization
    Birgin, E. G.
    Gardenghi, J. L.
    Martinez, J. M.
    Santos, S. A.
    [J]. OPTIMIZATION LETTERS, 2020, 14 (04) : 815 - 838
  • [8] THE USE OF QUADRATIC REGULARIZATION WITH A CUBIC DESCENT CONDITION FOR UNCONSTRAINED OPTIMIZATION
    Birgin, E. G.
    Martinez, J. M.
    [J]. SIAM JOURNAL ON OPTIMIZATION, 2017, 27 (02) : 1049 - 1074
  • [9] Worst-case evaluation complexity for unconstrained nonlinear optimization using high-order regularized models
    Birgin, E. G.
    Gardenghi, J. L.
    Martinez, J. M.
    Santos, S. A.
    Toint, Ph. L.
    [J]. MATHEMATICAL PROGRAMMING, 2017, 163 (1-2) : 359 - 368
  • [10] Worst-case evaluation complexity of regularization methods for smooth unconstrained optimization using Holder continuous gradients
    Cartis, C.
    Gould, N. I. M.
    Toint, Ph. L.
    [J]. OPTIMIZATION METHODS & SOFTWARE, 2017, 32 (06) : 1273 - 1298