ON THE GENERALIZED LANCZOS TRUST-REGION METHOD

被引:25
作者
Zhang, Lei-Hong [1 ,2 ]
Shen, Chungen [3 ]
Li, Ren-Cang [4 ]
机构
[1] Shanghai Univ Finance & Econ, Sch Math, 777 Guoding Rd, Shanghai 200433, Peoples R China
[2] Shanghai Univ Finance & Econ, Res Sch Interdisciplinary Sci, 777 Guoding Rd, Shanghai 200433, Peoples R China
[3] Univ Shanghai Sci & Technol, Coll Sci, Shanghai 200093, Peoples R China
[4] Univ Texas Arlington, Dept Math, Arlington, TX 76019 USA
基金
美国国家科学基金会;
关键词
trust-region subproblem; Lanczos method; conjugate gradient method; trust-region method; convergence; stopping criterion; CONJUGATE-GRADIENT METHOD; OPTIMIZATION; CONVERGENCE; SUBPROBLEMS;
D O I
10.1137/16M1095056
中图分类号
O29 [应用数学];
学科分类号
070104 ;
摘要
The so-called trust-region subproblem gets its name in the trust-region method in optimization and also plays a vital role in various other applications. Several numerical algorithms have been proposed in the literature for solving small-to-medium size dense problems as well as for large-scale sparse problems. The generalized Lanczos trust-region (GLTR) method proposed by [N. I. M. Gould, S. Lucidi, M. Roma and P. L. Toint, SIAM T. Opttm., 9 (1999), pp. 561-580] is a natural extension of the classical Lanczos method for the linear system to the trust-region subproblem. In this paper, we first analyze the convergence of GLTR to reveal its convergence behavior in theory and then propose new stopping criteria that can be integrated into GLTR for better numerical performance. Specifically, we develop a priori upper bounds for the convergence to both the optimal objective value as well as the optimal solution and argue that these bounds can be efficiently estimated numerically and serve as stopping criteria for iterative methods such as GLTR. Two sets of numerical tests are presented. In the first set, we demonstrate the sharpness of the upper bounds, and for the second set, we integrate the upper bound estimate into the Fortran routine GLTR in the library GALAHAD as new stopping criteria and test the trust-region solver TRU on the problem collection CUTEr. The numerical results show that, with the new stopping criteria in GLTR, the overall performance of TRU can be improved considerably.
引用
收藏
页码:2110 / 2142
页数:33
相关论文
共 37 条
  • [1] [Anonymous], 1997, Applied numerical linear algebra
  • [2] Bernstein S., 1912, Mem. Cl. Sci. Acad. R. Belg, V4, P1
  • [3] CHENEY E. W., 1982, INTRO APPROXIMATION
  • [4] CONN A. R., 2000, MOS SIAM SER OPTIM
  • [5] Benchmarking optimization software with performance profiles
    Dolan, ED
    Moré, JJ
    [J]. MATHEMATICAL PROGRAMMING, 2002, 91 (02) : 201 - 213
  • [6] COMPUTING OPTIMAL LOCALLY CONSTRAINED STEPS
    GAY, DM
    [J]. SIAM JOURNAL ON SCIENTIFIC AND STATISTICAL COMPUTING, 1981, 2 (02): : 186 - 197
  • [7] QUADRATICALLY CONSTRAINED LEAST-SQUARES AND QUADRATIC PROBLEMS
    GOLUB, GH
    VONMATT, U
    [J]. NUMERISCHE MATHEMATIK, 1991, 59 (06) : 561 - 580
  • [8] On solving trust-region and other regularised subproblems in optimization
    Gould N.I.M.
    Robinson D.P.
    Thorne H.S.
    [J]. Mathematical Programming Computation, 2010, 2 (01) : 21 - 57
  • [9] GALAHAD, a library of thread-safe Fortran 90 packages for large-scale nonlinear optimization
    Gould, NIM
    Orban, D
    Toint, PL
    [J]. ACM TRANSACTIONS ON MATHEMATICAL SOFTWARE, 2003, 29 (04): : 353 - 372
  • [10] Solving the trust-region subproblem using the Lanczos method
    Gould, NIM
    Lucidi, S
    Roma, M
    Toint, PL
    [J]. SIAM JOURNAL ON OPTIMIZATION, 1999, 9 (02) : 504 - 525