Abstract:The linear finite element approximation of a general linear diffusion problem with arbitrary anisotropic meshes is considered. The conditioning of the resultant stiffness matrix and the Jacobi preconditioned stiffness matrix is investigated using a density function approach proposed by Fried in 1973. It is shown that the approach can be made mathematically rigorous for general domains and used to develop bounds on the smallest eigenvalue and the condition number that are sharper than existing estimates in one … Show more
“…Bounds for the conditioning of the Schur complement are provided, e.g., in [13] or in Proposition 4.47 from [27] as function of the inf-sup constant and the condition number of the pressure mass matrix. A bound for the mass matrix with anisotropic elements is provided, e.g., in [28]. The results are consistent with the theory. )…”
“…Bounds for the conditioning of the Schur complement are provided, e.g., in [13] or in Proposition 4.47 from [27] as function of the inf-sup constant and the condition number of the pressure mass matrix. A bound for the mass matrix with anisotropic elements is provided, e.g., in [28]. The results are consistent with the theory. )…”
“…While using fine mesh throughout the whole domain may give an acceptable numerical solution, the computational cost of such an approach may be prohibitive. In contrast, a mesh that is coarse in one region and fine in another, while computationally more attractive, may be difficult to generate and is also known to cause difficulties with preconditioning [5,6,7,13]. Additionally, in some instances the boundary of the subregion may be geometrically complicated, or may change in time, requiring frequent remeshing or the use of complicated local refinement and derefinement techniques [12,15,22,23,24,27].…”
Problems with localized nonhomogeneous material properties arise frequently in many applications and are a well-known source of difficulty in numerical simulations. In certain applications (including additive manufacturing), the physics of the problem may be considerably more complicated in relatively small portions of the domain, requiring a significantly finer local mesh compared to elsewhere in the domain. This can make the use of a uniform mesh numerically unfeasible. While nonuniform meshes can be employed, they may be challenging to generate (particularly for regions with complex boundaries) and more difficult to precondition. The problem becomes even more prohibitive when the region requiring a finer-level mesh changes in time, requiring the introduction of refinement and derefinement techniques. To address the aforementioned challenges, we employ a technique related to the Fat boundary method [1,2,20] as a possible alternative. We analyze the proposed methodology from a mathematical point of view and validate our findings on two-dimensional numerical tests.
“…A more challenging benchmark test is to consider the FEM solution for a parameter field with non-smooth variation. In this case it is natural to anticipate that any significant jump discontinuities in the profile of p will have an adverse effect on the condition number of the stiffness matrix [13]. For our simulations we choose a piecewise constant approximation of the positive function p(x) .…”
We consider a sketched implementation of the finite element method for elliptic partial differential equations on high-dimensional models. Motivated by applications in real-time simulation and prediction we propose an algorithm that involves projecting the finite element solution onto a lowdimensional subspace and sketching the reduced equations using randomised sampling. We show that a sampling distribution based on the leverage scores of a tall matrix associated with the discrete Laplacian operator, can achieve nearly optimal performance and a significant speedup. We derive an expression of the complexity of the algorithm in terms of the number of samples that are necessary to meet an error tolerance specification with high probability, and an upper bound for the distance between the sketched and the high-dimensional solutions. Our analysis shows that the projection not only reduces the dimension of the problem but also regularises the reduced system against sketching error. Our numerical simulations suggest speed improvements of two orders of magnitude in exchange for a small loss in the accuracy of the prediction.1991 Mathematics Subject Classification. 65F05, 65M60, 68W20.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.