commons-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Luc Maisonobe (Resolved) (JIRA)" <j...@apache.org>
Subject [jira] [Resolved] (MATH-727) too large first step with embedded Runge-Kutta integrators (Dormand-Prince 8(5,3) ...)
Date Sat, 17 Dec 2011 16:57:30 GMT

     [ https://issues.apache.org/jira/browse/MATH-727?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Luc Maisonobe resolved MATH-727.
--------------------------------

    Resolution: Fixed

Fixed in subversion repository as of r1215524.
                
> too large first step with embedded Runge-Kutta integrators (Dormand-Prince 8(5,3) ...)
> --------------------------------------------------------------------------------------
>
>                 Key: MATH-727
>                 URL: https://issues.apache.org/jira/browse/MATH-727
>             Project: Commons Math
>          Issue Type: Bug
>    Affects Versions: 3.0
>            Reporter: Luc Maisonobe
>            Assignee: Luc Maisonobe
>             Fix For: 3.0
>
>
> Adaptive step size integrators compute the first step size by themselves if it is not
provided.
> For embedded Runge-Kutta type, this step size is not checked against the integration
range, so if the integration range is extremely short, this step size may evaluate the function
out of the range (and in fact it tries afterward to go back, and fails to stop). Gragg-Bulirsch-Stoer
integrators do not have this problem, the step size is checked and truncated if needed.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message