Zur numerischen Lösung des ersten biharmonischen Randwertproblems. (On the numerical solution of the first biharmonic boundary value problem) (Q5966455)

The system of equations resulting from a mixed finite element approximation of the first biharmonic boundary value problem is solved by various preconditioned Uzawa-type iterative methods. The preconditioning matrices are based on simple finite element approximations of the Laplace operator and some factorizations of the corresponding matrices. The most efficient variants of these iterative methods require asymptotically \(O(h^{-0.5}\ln \epsilon^{-1})\) iterations and \(O(h^{-p-0.5}\ln \epsilon^{-1})\) arithmetical operations only, where \(\epsilon\) denotes the relative accuracy and h is a mesh-size parameter such that the number of unknowns grows as \(O(h^{-p})\), \(h\to 0\). S.Sburlan: The author studies the first biharmonic boundary value problem by means of the mixed finite element approximation and iterative methods of Uzawa's type. The preconditioning matrices are based on simple finite element approximations of the Laplace operator which require symptotically \(O(h^{-0,5}\ln \epsilon^{-1})\) iterations and only \(O(h^{-p-0,5}\ln \epsilon^{-1})\) arithmetic operations; here \(\epsilon\) denotes the relative accuracy and h is the mesh-size parameter. Some numerical examples are also commented.