1.中国工程物理研究院研究生院,北京 100088
2.北京应用物理与计算数学研究所,北京 100094
3.中国工程物理研究院高性能数值模拟软件中心,北京 100088
段晓宇(1997年生),女;研究方向:数值并行算法;E-mail:duanxiaoyu19@gscaep.ac.cn
安恒斌(1974年生),男;研究方向:数值并行算法及应用;E-mail:an_hengbin@iapcm.ac.cn
纸质出版日期:2022-11-25,
网络出版日期:2022-05-07,
收稿日期:2021-11-19,
录用日期:2021-12-21
扫 描 看 全 文
段晓宇,安恒斌.Anderson加速方法中的松弛参数选取[J].中山大学学报(自然科学版),2022,61(06):113-122.
DUAN Xiaoyu,AN Hengbin.On choice of relaxation parameter for Anderson acceleration method[J].Acta Scientiarum Naturalium Universitatis Sunyatseni,2022,61(06):113-122.
段晓宇,安恒斌.Anderson加速方法中的松弛参数选取[J].中山大学学报(自然科学版),2022,61(06):113-122. DOI: 10.13471/j.cnki.acta.snus.2021B109.
DUAN Xiaoyu,AN Hengbin.On choice of relaxation parameter for Anderson acceleration method[J].Acta Scientiarum Naturalium Universitatis Sunyatseni,2022,61(06):113-122. DOI: 10.13471/j.cnki.acta.snus.2021B109.
Anderson加速方法是提高不动点迭代收敛速度的一类方法。该方法中的松弛参数一般在区间(0,1]内,且大多数选取为1。考虑扩大Anderson加速方法中松弛参数的选取范围,以求解非线性扩散方程的Picard迭代过程为例,比较了不同松弛参数选取方法的效果。数值结果表明,合适的松弛参数可以有效提高Anderson加速方法的效果。
Anderson acceleration method is a kind of method to improve the convergence speed of fixed-point iteration. The relaxation parameter is generally in the interval (0,1], and is selected as 1 in most cases. This paper expanded the selection range of the relaxation parameter. Taking the Picard iteration of solving the nonlinear diffusion equation as an example, the effect of the relaxation parameter selection method was compared. The numerical results show that the appropriate relaxation parameter can effectively improve the convergence speed of the Anderson acceleration method.
迭代方法Anderson加速非线性扩散方程参数选取
iterative methodsAnderson accelerationnonlinear diffusion equationsparameter selection
KNOLL D A, KEYES D E. Jacobian-free Newton-Krylov methods:A survey of approaches and applications [J]. Journal of Computational Physics, 2004, 193(2): 357-397.
PUGACHEV B P. Acceleration of the convergence of iterative processes and a method of solving systems of non-linear equations [J]. USSR Computational Mathematics and Mathematical Physics, 1977, 17(5): 199-207.
WYNN P. Acceleration techniques for iterated vector and matrix problems [J]. Mathematics of Computation, 1962, 16(79): 301-322.
BREZINSKI C, REDIVO-ZAGLIA M, SAAD Y. Shanks sequence transformations and Anderson acceleration [J]. SIAM Review, 2018, 60(3): 646-669.
JBILOU K, SADOK H. Vector extrapolation methods. Applications and numerical comparison [J]. Journal of Computational and Applied Mathematics, 2000, 122(1/2): 149-165.
BREZINSKI C. Recursive interpolation, extrapolation and projection [J]. Journal of computational and applied mathematics, 1983, 9(4): 369-376.
JBILOU K, SADOK H. Some results about vector extrapolation methods and related fixed-point iterations [J]. Journal of Computational and Applied Mathematics, 1991, 36(3): 385-398.
BREZINSKI C, REDIVO-ZAGLIA M. Extrapolation methods for the numerical solution of nonlinear Fredholm integral equations [J]. Journal of Integral Equations and Applications, 2019, 31(1): 29-57.
ANDERSON D G. Iterative procedures for nonlinear integral equations [J]. Journal of the ACM, 1965, 12(4): 547-560.
WALKER H F, NI P. Anderson acceleration for fixed-point iterations [J]. SIAM Journal on Numerical Analysis, 2011, 49(4): 1715-1735.
AN H, JIA X, WALKER H F. Anderson acceleration and application to the three-temperature energy equations [J]. Journal of Computational Physics, 2017, 347: 1-19.
SHI W, SONG S, WU H, et al. Regularized Anderson Acceleration for off-policy deep reinforcement learning [J]. Advances in Neural Information Processing Systems, 2019, 32: 10231-10241.
WALKER H F. Anderson acceleration: Algorithms and implementations [R]. MS-6-15-50, Worcester Polytechnic Institute Mathematical Sciences Department, 2011.
POTRA F A, ENGLER H. A characterization of the behavior of the Anderson acceleration on linear problems [J]. linear Algebra and its Applications, 2013, 438(3): 1002-1011.
EVANS C, POLLOCK S, REBHOLZ L G, et al. A proof that Anderson acceleration improves the convergence rate in linearly converging fixed-point methods (but not in those converging quadratically) [J].SIAM Journal on Numerical Analysis, 2020, 58(1): 788-810.
FANG H, SAAD Y. Two classes of multisecant methods for nonlinear acceleration [J]. Numerical Linear Algebra with Applications, 2009, 16(3): 197-221.
POLLOCK S, REBHOLZ L G. Anderson acceleration for contractive and noncontractive operators [J]. IMA Journal of Numerical Analysis, 2021, 41(4): 2841-2872.
AKSENOV V, CHERTOV M, SINKOV K. Application of accelerated fixed-point algorithms to hydrodynamic well-fracture coupling [J]. Computers and Geotechnics, 2021, 129: 103783.
YANG Y. Anderson acceleration for seismic inversion [J]. Geophysics, 2021, 86(1): R99-R108.
BIAN W, CHEN X, KELLEY C T. Anderson acceleration for a class of nonsmooth fixed-point problems [J]. SIAM Journal on Scientific Computing, 2021,43(5): S1-S20.
FILIPPINI M, ALOTTO P, GIUST A. Anderson acceleration for electromagnetic nonlinear problems [J]. COMPEL-the International Journal for Computation and Mathematics in Electrical and Electronic Engineering, 2019,38(5):1493-1506.
PENG Y, DENG B, ZHANG J, et al. Anderson acceleration for geometry optimization and physics simulation [J]. ACM Transactions on Graphics, 2018, 37(4): 1-14.
LOTT P A, WALKER H F, WOODWARD C S, et al. An accelerated Picard method for nonlinear systems related to variably saturated flow [J]. Advances in Water Resources, 2012, 38: 92-101.
秦子康, 安恒斌, 王新玉. 两类向量序列加速收敛方法比较[J]. 数值计算与计算机应用, 2021, 42(4): 379-394.
0
浏览量
1
下载量
0
CSCD
关联资源
相关文章
相关作者
相关机构