On the equivalence of extrapolation and Richardson's iteration and its applications (Q556888): Difference between revisions

In 1984 two articles [\textit{G. Opfer} and \textit{G. Schober}, ``Richardson's iteration for nonsymmetric matrices'', ibid. 58, 343--361 (1984; Zbl 0565.65012); \textit{A. Hadjidimos}, ``The Optimal Solution to the Problem of Complex Extrapolation of a First Order Scheme'', ibid. 62, 241--261 (1984; Zbl 0567.65015)] appeared in the same journal solving the problem of optimizing the convergence factor of the stationary Richardson's iteration (i.e., \(x^{(m+1)} = x^{(m)} -a (Ax^{(m)} - b), m=0, 1, 2, \cdots \)) for the solution of the system \(Ax=b, A \in C^{n,n} ,\) det(A) \(\!=0, b \in C^n .\) Opfer and Schober determined the optimal parameter \(a\), when the spectrum of \(A\) is contained in a certain compact set \(S_1 \) not including the origin. Hadjidimos considered, for the solution of \(Ax=b\), the consistent linear scheme \(x^{(m+1)} = Tx^{(m)} +c, m=0, 1, 2, \dots ,\) for which the spectrum of \(T\) is contained in a certain compact convex set \(S_2\) not including the point 1, and determined the optimal extrapolation parameter \(\omega\) of its extrapolation scheme (i.e. \(x^{(m+1)} = ((1-\omega)I + \omega T)x^{(m)} + \omega c, m=0, 1, 2, \cdots \)). Because the approach in obtaining the optimal parameters in the two articles was quite different it seems it is difficult for any one to notice their equivalence and, therefore, to use them in trivial cases. In this article the author shows the equivalence of the optimal results of Opfer and Schober, and of Hadjidimos, explores them further, obtains new results and gives applications for cases which may occur in practice. Finally, based on observations regarding the optimal two-cyclic successive overrelaxation and accelerated overrelaxation (AOR) methods in two real case problems, he solves the problem of an ``optimal'' two-cyclic AOR in the more general complex case.