Since at the end of each iteration we have an interval known to
contain the solution, and since we choose the middle of this interval
(
-wise), we can use its half-size as a convergence criterion.
We stop the iterations when:
|
(14) |
 |
where
is the convergence threshold.
When this threshold is met, we use
and we
compute
using the following expression which is true when
has been found:
|
(15) |
 |
Luc Maisonobe
2006-02-04