Vector Aitken's delta-square accelerator

Vector version of Aitken's accelerator

Let the result of an iterative process be a vector. The solution may be reached by the Irons-Tuck version of the Aitken's delta-square process. Since the original algorithm may suffer by a numerical instability, it can be stabilized by employing a built-in weighing function fw depending on the current order of iteration.

A problem of finding maximum eigenvalue of the Hilbert matrix by the power method is implemented by the script demoAitkd2.