×

Optimal finite element mesh for elliptic equation of divergence form. (English) Zbl 1069.65134

The authors discuss the generation of triangular meshes for elliptic partial differential equations that result in stiffness matrices with minimal condition numbers. The condition number \(\mathcal K\) of the stiffness matrix for the Dirichlet problem is estimated in terms of the elemental quantities \({\mathcal K}_e\), defined as the ratio of the largest eigenvalue of the elemental stiffness matrix divided by the smallest eigenvalue of the elemental mass matrix. Only linear finite elements on triangles are considered.
A direct calculation for the Poisson equation shows that the quantity \({\mathcal K}_e\) is minimized for equilateral triangles. For right triangles, it is minimized for isosceles right triangles. For a general coercive elliptic equation with coefficient matrix \(A\), the authors show that the metric induced by the matrix \(A^{-1/2}\) should be used. When the coefficient matrix varies with position, optimality can be approximated using the value of \(A\) at each element’s centroid. These results give rise to a mesh generation algorithm based on generating equal-legged triangles in the \(A^{-1/2}\) metric.
Although no proof is presented that this mesh strategy results in the minimal condition number \(\mathcal K\), a direct calculation for the case of a uniform isosceles right triangular mesh for the Poisson equation shows that \(\mathcal K\) agrees with each \({\mathcal K}_e\), supporting the likelihood that the strategy does minimize \(\mathcal K\). In addition, several numerical examples are presented.

MSC:

65N50 Mesh generation, refinement, and adaptive methods for boundary value problems involving PDEs
65N30 Finite element, Rayleigh-Ritz and Galerkin methods for boundary value problems involving PDEs
35J05 Laplace operator, Helmholtz equation (reduced wave equation), Poisson equation
35J25 Boundary value problems for second-order elliptic equations
65F35 Numerical computation of matrix norms, conditioning, scaling
PDFBibTeX XMLCite
Full Text: DOI

References:

[1] Atkinson, K. E., An Introduction to Numerical Analysis (1988), John Wiley & Sons
[2] Axelsson, O.; Barker, V. A., Finite Element Solution of Boundary Value Problems (1984), Academic Press · Zbl 0537.65072
[3] Babuška, I.; Craig, A.; Mandel, J.; Pitkäranta, J., Efficient preconditioning for the \(p\)-version finite element method in two dimensions, SIAM J. Numer. Anal, 28, 624-661 (1991) · Zbl 0754.65083
[4] Bank, R. E.; Scott, L. R., On the conditioning of finite element equations with highly refined meshes, SIAM J. Numer. Anal, 26, 1383-1394 (1989) · Zbl 0688.65062
[5] Buscaglia, G. C.; Dari, E. A., Anisotropic mesh optimization and its application in adaptivity, Int. J. Numer. Mech. Eng, 40, 4119-4136 (1997) · Zbl 0899.76264
[6] Ciarlet, P. G., The Finite Element Method for Elliptic Problems (1978), North-Holland: North-Holland Amsterdam · Zbl 0445.73043
[7] Fried, I., Condition of finite element stiffness matrices generated from non-uniform meshes, AIAA J, 10, 219-221 (1972) · Zbl 0242.65046
[8] Fried, I., Bounds on the spectral and maximum norms of the finite element stiffness, flexibility and mass matrices, Int. J. Solids Struct, 9, 1013-1034 (1973) · Zbl 0263.73049
[9] Golub, G.; Van Loan, C., Matrix Computations (1983), Johns Hopkins Press: Johns Hopkins Press Baltimore · Zbl 0559.65011
[10] Hu, N.; Guo, X.-Z.; Katz, I. N., Bounds for eigenvalues and condition numbers in the \(p\)-version of the finite element method, Math. Comput, 67, 1423-1450 (1998) · Zbl 0907.65112
[11] Jost, J., Riemannian Geometry and Geometric Analysis (1995), Springer-Verlag: Springer-Verlag Heidelberg · Zbl 0828.53002
[12] Maitre, J.-F.; Pourquier, O., Condition number and diagonal preconditioning: comparison of the \(p\)-version and the spectral element methods, Numer. Math, 74, 69-84 (1996) · Zbl 0858.65114
[13] Olsen, E. T.; Douglas, J., Bounds on spectral condition number of matrices arising in the \(p\)-version of the finite element method, Numer. Math, 69, 333-352 (1995) · Zbl 0834.65034
[14] Ramage, A.; Wathen, A. J., On preconditioning for the finite element equations on irregular grids, SIAM J. Matrix Anal. Appl, 15, 909-921 (1994) · Zbl 0807.65049
[15] Vasilevskii, Y. V.; Lipnikov, K. N., An adaptive algorithm for quasioptimal mesh generation, Comput. Math. Math. Phys, 39, 1532-1551 (1999) · Zbl 0981.65141
[16] Young, D. M., Iterative Solution of Large Linear Systems (1971), Academic Press · Zbl 0204.48102
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.