Dynamic iteration methods applied to linear DAE systems (Q1118990): Difference between revisions

The author considers the iterative solution of a differential-algebraic equation (*) \(B\dot x+Ax=f\), \(x(0)=x_ 0\) with constant matrices A, B by iteration methods of the form \(M_ B\dot x^ k+M_ Ax^ k=N_ B\dot x^{k-1}+N_ Ax^{k-1}+f\), \(x^ k(0)=x_ 0\) where \(A=M_ A-N_ A\), \(B=M_ B-N_ B\) are suitable splittings. Necessary conditions for the convergence of the splittings are given which are brought by the singularity of B into consideration. In the case (**) A positive definite and B positive semidefinite convergence results for the iteration are given. The present paper continues a recent corresponding investigation by the author and \textit{O. Nevanlinna} [SIAM J. Sci. Stat. Comp. 8(4), 459-482 (1987; Zbl 0625.65063)] for the case \(B=I.\) Remark: The assumption made in Theorem 6 that (*) has index 1 is superfluous since it is a consequence of the further assumptions (**).