摘要
Anderson算法是求解非线性方程组的有效加速迭代方法。本文采用Anderson(m,β)算法求解二维和三维Burgers方程的Crank-Nicolson格式离散所得的非线性方程组。数值计算结果表明,当算法参数β=-0.5时,由离散所得的非线性方程组的Anderson迭代解的收敛性达到最优。
Anderson acceleration is an effective accelerating iterative method to solve nonlinear equations. Anderson acceleration( m,β) was used to solve nonlinear equations resulting from the discretization by Crank-Nicolson schemes for two and three dimensional coupled Burgers' equations. The numerical results indicate that when,convergence of iterative solutions for nonlinear equations is optimal in algorithm Anderson( m,β).
引文
[1]FRIED JJ,COMBARNOUS M.Dispersion in porousmedia[J].Advance in Hydroscience,1971,7(1):169-282.
[2]FLETCHER C A J.A comparison of finite element and finite difference of the one-and two-dimensional Burgers'equations[J].Journal of Computational Physics,1983,51(1):159-188.
[3]JAIN P C.Numerical solution of coupled Burgersequations[J].International Journal of Non-Linear Mechanics,1978,13(4):213-222.
[4]GOYON O.Multilevel schemes for solving unsteadyequations[J].International Journal for Numerical Methods in Fluids,1996,22(10):937-959.
[5]KWEYU M C.Crank-Nicolson scheme for numerical solutions of two-dimensional coupled Burgersequations[J].International Journal of Scientific&Engineering Research,2011,2(5):1-6.
[6]SRIVASTAVA V K,TAMSIR M,BHARDWAJ U.Crank-Nicolson Scheme for Numerical Solutions of Two-dimensional Coupled Burgers'Equations[J].International Journal of Scientific&Engineering Research,2011,2(5):1-7.
[7]CAMPOS M D,ROMAO E C.A high-order finite-difference scheme with a linearization technique for solving of three-dimensional Burgersequation[J].Computer Modeling In Engineering and Sciences,2014,103(3):139-154.
[8]BAHADIR A R.A fully implicit finite-difference scheme for two-dimensional Burgers'equations[J].Applied Mathematics and Computation,2003,137(1):131-137.
[9]SRIVASTAVA V K,AWASTHI M K,SINGH S.An implicit logarithmic finite-difference technique for two dimensional coupled viscous Burgers'equation[J].AIP Advances,2013,3(12):122105-122114.
[10]KELLEY C T.Iterative Methods for Linear and Nonlinear Equations[M].Philadelphia:SIAM,1995.
[11]KNOLL D,KEYES D.Jacobian-free Newton-Krylov methods:A survey of approaches and applications[J].Journal of Computational Physics,2004,193(2):357-397.
[12]ANDERSON D G.Iterative procedures for nonlinear integralequations[J].Journal of The ACM,1965,12(4):547-560.
[13]CARLSON NN,MILLER K.Design and application of a gradientweighted moving finite element code I:In one dimension[J].SIAM Journal on Scientific Computing,1998,19(3):728-765.
[14]BROYDEN C G.A class of methods for solving nonlinear simultaneousequations[J].Mathematics of Computation,1965,19(2):577-593.
[15]CALEF M,FICHTL E,WARSA J,et al.Nonlinear Krylov acceleration applied to a discrete orinates formulation of the k-eigenvalue problem[J].Journal of Computational Physics,2013,238(1):188-209.
[16]WILLERT J,TAITANO W T,KNOLL D.Leveraging Anderson acceleration for improved convergence of iterative solutions to transportsystems[J].Journal of Computational Physics,2014,273(15):278-286.
[17]WALKER H F,Ni P.Anderson acceleration for fixed-pointiterations[J].SIAM Journal on Numerical Analysis,2011,49(4):1715-1735.
[18]TOTH A,KELLEY C T.Convergence analysis for Andersonacceleration[J].SIAM Journal on Numerical Analysis,2015,53(2):805-819.
[19]ASLAN Y.The semi-approximate approach for solving Burgers'equation with high Reynoldsnumber[J].Applied Mathematics and Computation,2005,163(1):131-145.