SUPERCONVERGENCE OF A DISCONTINUOUS GALERKIN METHOD FOR FRACTIONAL DIFFUSION AND WAVE EQUATIONS

被引:157
作者
Mustapha, Kassem [1 ]
Mclean, William [2 ]
机构
[1] King Fahd Univ Petr & Minerals, Dept Math & Stat, Dhahran 31261, Saudi Arabia
[2] Univ New S Wales, Sch Math & Stat, Sydney, NSW 2052, Australia
关键词
finite elements; dual problem; postprocessing; FINITE-DIFFERENCE METHOD; IMPLICIT NUMERICAL-METHOD; EVOLUTION EQUATION; ANOMALOUS DIFFUSION; ORDER; MEMORY; DISCRETIZATION; STABILITY; ACCURACY; SCHEMES;
D O I
10.1137/120880719
中图分类号
O29 [应用数学];
学科分类号
070104 ;
摘要
We consider an initial-boundary value problem for partial derivative(t)u - partial derivative(-alpha)(t)del(2)u - f(t), that is, for a fractional diffusion (-1 < alpha < 0) or wave (0 < alpha < 1) equation. A numerical solution is found by applying a piecewise-linear, discontinuous Galerkin (DG) method in time combined with a piecewise-linear, conforming finite element method in space. The time mesh is graded appropriately near t = 0, but the spatial mesh is quasi-uniform. Previously, we proved that the error, measured in the spatial L-2-norm, is of order k(2+alpha-) + h(2)l(k), uniformly in t, where k is the maximum time step, h is the maximum diameter of the spatial finite elements, alpha(-) = min(alpha,0) <= 0, and l(k) = max(1, vertical bar log k vertical bar). Here, we prove convergence of order k(3+2 alpha-) l(k) + h(2) at each time level t(n) for -1 < alpha < 1. Thus, if -1/2 < alpha < 1, then the DG solution is superconvergent, which generalizes a known result for the classical heat equation (i.e., the case alpha = 0). A simple postprocessing step employing Lagrange interpolation leads to superconvergence for any t. Numerical experiments indicate that our theoretical error bound is pessimistic if alpha < 0. Ignoring logarithmic factors, we observe that the error in the DG solution at t = t(n), and after postprocessing at all t, is of order k(3+alpha-) + h(2) for -1 < alpha < 1.
引用
收藏
页码:491 / 515
页数:25
相关论文
共 45 条
[11]   Wave propagation in media with singular memory [J].
Hanyga, A .
MATHEMATICAL AND COMPUTER MODELLING, 2001, 34 (12-13) :1399-1421
[12]   ERROR ESTIMATES FOR A SEMIDISCRETE FINITE ELEMENT METHOD FOR FRACTIONAL ORDER PARABOLIC EQUATIONS [J].
Jin, Bangti ;
Lazarov, Raytcho ;
Zhou, Zhi .
SIAM JOURNAL ON NUMERICAL ANALYSIS, 2013, 51 (01) :445-466
[13]  
Kilbas AA, 2006, THEORY APPL FRACTION
[14]   The accuracy and stability of an implicit solution method for the fractional diffusion equation [J].
Langlands, TAM ;
Henry, BI .
JOURNAL OF COMPUTATIONAL PHYSICS, 2005, 205 (02) :719-736
[15]   Numerical method and analytical technique of the modified anomalous subdiffusion equation with a nonlinear source term [J].
Liu, F. ;
Yang, C. ;
Burrage, K. .
JOURNAL OF COMPUTATIONAL AND APPLIED MATHEMATICS, 2009, 231 (01) :160-176
[16]   On the numerical inversion of the Laplace transform of certain holomorphic mappings [J].
López-Fernández, M ;
Palencia, C .
APPLIED NUMERICAL MATHEMATICS, 2004, 51 (2-3) :289-303
[17]   A spectral order method for inverting sectorial Laplace transforms [J].
Lopez-Fernandez, Maria ;
Palencia, Cesar ;
Schaedle, Achim .
SIAM JOURNAL ON NUMERICAL ANALYSIS, 2006, 44 (03) :1332-1350
[18]  
Mainardi F, 2001, J COMPUT ACOUST, V9, P1417, DOI 10.1016/S0218-396X(01)00082-6
[19]  
Mathai AM, 2010, H-FUNCTION: THEORY AND APPLICATIONS, P1, DOI 10.1007/978-1-4419-0916-9
[20]   Discretization with variable time steps of an evolution equation with a positive-type memory term [J].
McLean, W ;
Thomee, V ;
Wahlbin, LB .
JOURNAL OF COMPUTATIONAL AND APPLIED MATHEMATICS, 1996, 69 (01) :49-69