Weak-convergence theory of quasi-nonnegative splittings for singular matrices. (Q1412331): Difference between revisions

The authors present a new concept for a \(k\)-order quasi-nonnegative splitting. They consider large sparse system of linear equations \(Ax=b\) where the coefficient matrix \(A \in \mathbb{C}^{n \times n}\) is singular and the right-hand-side vector \(b \in R(A)\) where \(R(A)\) denotes the range of the matrix \(A\). In the case where this matrix can be split as \(A=M-N\) with \(M\) being nonsingular the authors construct a splitting iterative method with an iteration matrix \(T=M^{-1}N\). The weak-convergence of the constructed iterative method is studied. The equivalence among several conditions for the so called \textit{k}-order weak-convergence is established. Further, the monotone convergence properties of the \textit{k}-order quasi-nonnegative splittings are studied and discussed. The method and the convergence theory developed is illustrated with a simple numerical example.