[
https://issues.apache.org/jira/browse/MATH-937?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13590726#comment-13590726
]
Luc Maisonobe commented on MATH-937:
------------------------------------
OK, I'll do the javadoc change with the link you suggest.
I'm not sure about ignoring the exception, so if you want to give it a try,
I'll review your proposal.
If we go along this way, we should take care to ignore *only* this specific
exception, as there can be
many other reasons for which we fail to find the root of arbitrary functions
provided by user, and such
failures should be reported.
The call is already set up to force bracketing, and as you noticed, when it
fails to do so there is a
reason for failure. In this case, it was because we found the same event again
because the function was
"almost" continuous (the discontinuity was at computing precision level). Here,
ignoring the error and
continuing would have been the way to go. I'm not sure this would always be the
case, and I'm not sure
either that the algorithm can recover a sane state after this occurs. So we
have to see what you propose
befre deciding to implement it or not.
> NoBracketingException after event was found
> -------------------------------------------
>
> Key: MATH-937
> URL: https://issues.apache.org/jira/browse/MATH-937
> Project: Commons Math
> Issue Type: Bug
> Affects Versions: 3.1.1
> Reporter: Christoph Höger
> Attachments: ApacheCommonsBouncingBall.java,
> ApacheCommonsBouncingBallTest.java
>
>
> The BracketingNthOrderBrentSolver used by the EmbeddedRungeKuttaIntegrator
> fails, if an event is detected twice with a NoBracketingException.
> The problem lies in line EventState.java line 262 (version 3.1.1). Here the
> event detection function f is applied to an arbitrary choosen value of time.
> If the event detector crosses zero before this time, the solver throws the
> mentioned exception.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
