• 论文 • 上一篇    

EIM魔数点特性研究及其最小二乘格式

龚禾林1, 张世全2, Yvon Maday3   

  1. 1. 上海交通大学巴黎卓越工程师学院, 上海 200240;
    2. 四川大学数学学院, 成都 610064;
    3. 法国索邦大学LJLL实验室, 巴黎 70005, 法国
  • 收稿日期:2021-10-31 发布日期:2023-03-16
  • 基金资助:
    国家自然科学基金(11905216)资助.

龚禾林, 张世全, Yvon Maday. EIM魔数点特性研究及其最小二乘格式[J]. 数值计算与计算机应用, 2023, 44(1): 25-36.

Gong Helin, Zhang Shiquan, Yvon Maday. THE OPTIMUM OF EIM MAGIC POINTS AND THE LEAST-SQUARES FORM[J]. Journal on Numerica Methods and Computer Applications, 2023, 44(1): 25-36.

THE OPTIMUM OF EIM MAGIC POINTS AND THE LEAST-SQUARES FORM

Gong Helin1, Zhang Shiquan2, Yvon Maday3   

  1. 1. Paris Elite Institute of Technology, Shanghai Jiao Tong University, Shanghai 200240, China;
    2. School of Mathematics, Sichuan University, Chengdu 610064, China;
    3. Institut Universitaire de France;Sorbonne Universites, UPMC Univ Paris 06 and CNRS UMR 7598, Laboratoire Jacques-Louis Lions, Paris, F-75005, France
  • Received:2021-10-31 Published:2023-03-16
经验插值法(empirical interpolation method,EIM)首先由Yvon Maday和他的合作者在2004年提出,旨在提升非仿射或非线性偏微分方程模型降阶(reduced basis technique)的计算效率,随后在模型降阶和数据同化领域得到了广泛应用.EIM的计算过程分为离线、在线两个过程:离线阶段,基于待插值函数空间的大量样本函数,通过EIM算法逐一计算插值基函数和插值点(魔数点);在线阶段,基于在魔数点的函数值和基函数,在线重构待插值函数.本文重点研究了EIM算法得到的插值点的空间分布特性,提出了最小二乘格式的EIM (LS-EIM)以进一步提升EIM精度和稳定性.比较了EIM算法确定的魔数点和其它各种采样方法确定的点对LS-EIM的收敛性和精度的影响.通过数值计算发现,相比随机采样和其他方法选取的采样点,EIM算法得到的魔数点用于LS-EIM可获得最快收敛速度和最优重构精度,通常仅需不到2倍于基函数维数$n$的魔数点数,即不到2$n$个魔数点,LS-EIM即可实现对最佳重构的逼近.
The empirical interpolation method (EIM) has been introduced to extend the reduced basis technique to nonaffine and nonlinear partial differential equations (PDEs), and then widely used in the field of model order reduction and data assimilation. The efficiency of EIM is achieved through an offline-online decomposition procedure. In the offline phase, a set of the snapshots (manifold) of the underlying function are solved, and then a set of basis and interpolation points (magic point) are solved recursively based on the manifold. In the online phase, the interpolation to an unknown function is done by combining the observations of the function itself at some points and the reduced basis. In this paper, the performance of the magic points is analyzed. To further improve the accuracy and stability, we propose a least-squares framework of EIM (LS-EIM) with more points than originally required, and different kind of points sampling methods are investigated. Our numerical finding is that, with only two times of measurements from the first n EIM magic points, LS-EIM allows to reach an accuracy similar to that of best approximation in the reduce space which is spanned by the first EIM basis functions, with a high stability performance. Furthermore, these additional EIM points are better than a choice of a random sampling or other sampling methods.

MR(2010)主题分类: 

()
[1] Barrault M, Maday Y, Nguyen N C, Patera AT. An ‘empirical interpolation’ method:application to efficient reduced-basis discretization of partial differential equations[J]. Comptes Rendus Mathematique, 2004, 339(9):667-672.
[2] Grepl M A, Maday Y, Nguyen N C, Patera A T. Efficient reduced-basis treatment of nonaffine and nonlinear partial differential equations[J]. ESAIM:Mathematical Modelling and Numerical Analysis, 2007, 41(3):575-605.
[3] Maday Y, Nguyen N C, Patera A T, Pau S H. A general multipurpose interpolation procedure:the magic points[J]. Communications on Pure & Applied Analysis, 2009, 8(1):383-404.
[4] Maday Y, Patera A T, Turinici G. A priori convergence theory for reduced-basis approximations of single-parameter elliptic partial differential equations[J]. Journal of Scientific Computing, 2002, 17(1-4):437-446.
[5] Maday Y. Reduced basis method for the rapid and reliable solution of partial differential equations[C]. In International Congress of Mathematicians, Eur. Math. Soc. Zurich. Citeseer. 2006, III:1255-1270.
[6] Rozza G, Huynh D B P, Patera A T. Reduced basis approximation and a posteriori error estimation for affinely parametrized elliptic coercive partial differential equations[J]. Archives of Computational Methods in Engineering, 2008, 15(3):229-275.
[7] Maday Y, Mula O, Turinici G. Convergence analysis of the generalized empirical interpolation method[J]. SIAM Journal on Numerical Analysis, 2016, 54(3):1713-1731.
[8] Argaud J P, Bouriquet B, De Caso F, Gong H, Maday Y, Mula O. Sensor placement in nuclear reactors based on the generalized empirical interpolation method[J]. Journal of Computational Physics, 2018, 363:354-370.
[9] Gong H, Argaud J P, Bouriquet B, Maday Y. The Empirical Interpolation Method applied to the neutron diffusion equations with parameter dependence[C]. In Proceedings of PHYSOR. 2016.
[10] Argaud J P, Bouriquet B, Gong H, Maday Y, Mula O. Stabilization of (G) EIM in presence of measurement noise:application to nuclear reactor physics[C]. Spectral and High Order Methods for Partial Differential Equations ICOSAHOM 2016. Springer, Cham, 2017. 133-145.
[11] Ghavamian F, Tiso P, Simone A. POD-DEIM model order reduction for strain-softening viscoplasticity[J]. Computer Methods in Applied Mechanics and Engineering, 2017, 317:458-79.
[12] Zhou Y B. Model Reduction for Nonlinear Dynamical Systems with Parametric Uncertainties[D]. Massachusetts Institute of Technology, 2012.
[13] Everson R, Sirovich L. Karhunen-Loeve procedure for gappy data[J]. Journal of the Optical Society of America A, 1995, 12(8):1657-1664.
[14] Willcox K E. Unsteady flow sensing and estimation via the gappy proper orthogonal decomposition[J]. Computers & fluids, 2006, 35(2):208-226.
[15] Bebendorf M, Maday Y, Stamm B. Comparison of some reduced representation approximations. In Reduced Order Methods for Modeling and Computational Reduction[B]. Springer, 2014, 67-100.
[16] Cohen A, Davenport M A, Leviatan D. On the stability and accuracy of least squares approximations[J]. Foundations of computational mathematics, 2013, 13(5):819-834.
[17] Cohen A, Migliorati G. Optimal weighted least-squares methods[J]. The SMAI journal of computational mathematics, 2017, 3:181-203.
[18] Nevai P, Freud G. Orthogonal polynomials and Christoffel functions:A case study[J]. Journal of Approximation Theory, 1986, 48(1):3-167.
[1] 金鑫, 汪韬. 两阶段Arrhenius方程参数优化算法[J]. 数值计算与计算机应用, 2022, 43(3): 329-342.
[2] 兰林, 朱春钢. 基于Newton迭代法的最小二乘渐进迭代逼近[J]. 数值计算与计算机应用, 2022, 43(1): 88-111.
[3] 王丽, 苗真, 蒋耀林. 变系数偏微分方程的Galerkin KPOD模型降阶方法[J]. 数值计算与计算机应用, 2021, 42(3): 226-236.
[4] 牛善洲, 刘宏, 朱赟, 喻高航, 马建华. 基于广义惩罚加权最小二乘的低剂量CT重建方法[J]. 数值计算与计算机应用, 2021, 42(3): 289-302.
[5] 陈世军, 卢民荣. 单变量矩阵方程子矩阵约束下牛顿-MCG算法[J]. 数值计算与计算机应用, 2020, 41(4): 306-314.
[6] 王丽. 三类新的求解广义最小二乘问题的预处理GAOR方法[J]. 数值计算与计算机应用, 2020, 41(4): 282-296.
[7] 徐正伟, 王皓, 胡兵, 张世全. 基于非线性最小二乘的Arrhenius方程参数估计[J]. 数值计算与计算机应用, 2019, 40(4): 279-290.
[8] 张舰齐, 王丽琼, 左瑞亭. 区域气候数值模式预测误差的非线性预估及订正应用试验[J]. 数值计算与计算机应用, 2017, 38(1): 11-25.
[9] 乐航睿, 杨庆之. 求解正则化最小二乘问题的一个非精确交替方向乘子法[J]. 数值计算与计算机应用, 2016, 37(3): 223-232.
[10] 邓远北, 刘莹, 杨娟. 几种特殊循环矩阵的PROCRUSTES问题[J]. 数值计算与计算机应用, 2016, 37(1): 11-24.
[11] 杨平, 徐康丽, 蒋耀林. K-power系统在 Grassmann流形上的单侧模型降阶方法[J]. 数值计算与计算机应用, 2016, 37(1): 41-56.
[12] 陈春岳, 蒋耀林. 时域投影模型降阶方法的小样本误差估计[J]. 数值计算与计算机应用, 2012, 33(4): 283-292.
[13] 袁占斌, 聂玉峰, 欧阳洁. 基于泰勒基函数的移动最小二乘法及误差分析[J]. 数值计算与计算机应用, 2012, (1): 25-31.
[14] 童乔凌, 刘天桢, 童恒庆. 结构方程模型的约束最小二乘解与确定性算法[J]. 数值计算与计算机应用, 2009, 30(3): 170-180.
[15] 袁飞, 张凯院. 矩阵方程AXB + CXTD=F自反最小二乘解的迭代算法[J]. 数值计算与计算机应用, 2009, 30(3): 195-201.
阅读次数
全文


摘要