In this paper we present a new incomplete LU decomposition which is based on an existing sparse direct solver. In contrast to many incomplete LU decompositions this ILU incorporates information about the inverse factors L−1 and U−1 which have direct influence on the dropping strategy. We demonstrate in several large scale examples that this implementation constructs a robust preconditioner.

In this paper three iterative methods are studied: preconditioned GMRES with ILU preconditioning, GMRESR with multigrid as inner loop and multigrid for the solution of the incompressible Navier-Stokes equations in general coordinates. Robustness and e ciency of the three methods are investigated and compared. Numerical results show that the second method is very promising.

A combined strategy of numerical dropping and ll-in is described for constructing an eecient ILU preconditioner which is well-adapted to Jacobi matrices generated by reaction-diiusion equations modeling diiusion processes in semiconductors. The method is based on a physically motivated subdivision of un-knowns into "essential" and "non-essential" ones.

in this paper, we consider an arbitrary binary polynomial sequence {a_n} and then give a lower triangular matrix representation of this sequence. as main result, we obtain a factorization of the innite generalized pascal matrix in terms of this new matrix, using a riordan group approach. further some interesting results and applications are derived.

in this paper‎, ‎we determine the simple groups $g=ab$‎, ‎where $b$ is isomorphic to $l_{3}(4)$ and $a$ isomorphic to an alternating or a symmetric group on $ngeq5$‎, ‎letters‎.

In the present study, an advection-diffusion problem has been considered for the numerical solution. The continuum equation is discretized using both upwind and centered scheme. The linear system is solved using the ILU preconditioned BiCGSTAB method. Both Dirichlet and Neumann boundary condition has been considered. The obtained results have been compared for different cases.