A generalized ADI iterative method (Q1326479)

The ADI iterative method for the solution of Sylvester's equation \(AX - XB = C\) proceeds by strictly alternating between the solution of the two equations \((A - \delta_{k+1} I)X_{2k+1} = X_{2k}(B - \delta_{k + 1} I)+ C\) and \(X_{2k + 2}(B - \tau_{k + 1} I) = (A - \tau_{k + 1} I) X_{2k+1} - C\), for \(k = 0,1,2,\dots\). The determination of optimal parameters leads to a problem of a type which is often encountered in rational approximation and which has been already studied by Zolotarov. That is: given two disjoint sets \(F\) and \(G\) in the complex plane and find a rational function such that the term \(\max_{z \in F}| r(z)| / \min_{z \in G} | r(z)|\) is as small as possible. The solution is also related to potential theory. The authors emphasize that a strict alternation is not always the best. - - The list of references contains 47 entries.