Errata to: ``Elimination and fixed point iterations'' (Q1324335): Difference between revisions

A programming mistake led to numerical errors in Tables 4.1 and 4.2 of the authors' paper [ibid. 25, No. 5, 43-53 (1993; Zbl 0780.65030)]. The two leftmost columns corresponding to the Newton iterations and their residuals in both tables are correct; on the other hand, those corresponding to the Newton-Fourier iterations are wrong. The calculations have been done this time in double precision but using a PC. Before elimination, the Newton iterates converge in six steps, whereas the Newton-Fourier iterates need eight steps; Table 4.1 contains the first five iterates and their residuals. After elimination, the former iterates converge in five steps, while the latter need six, as is shown in Table 4.2. (See revised tables on the following pages).