[
https://issues.apache.org/jira/browse/MATH-1137?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14064358#comment-14064358
]
Nigel Goodwin commented on MATH-1137:
-------------------------------------
My suspicion is being directed towards
trustRegionCenterInterpolationPointIndex
In some cases this seems to be zero based, in others 1 based.
I was looking at the values of k in the loop and the values of the test at
about line 790. The skipped k values were not consistent between Fortran and
Java, even when knew was set the same at line 790....
Also KOPT v KNEW looks suspicious - the Fortran has a test against KOPT in
ALTMOV, whereas the java has a test against knew...
This is what I mean about the long hard slog comparing Java and Fortran. Unit
tests are for lazy people!
> BOBYQA incorrect indexing
> -------------------------
>
> Key: MATH-1137
> URL: https://issues.apache.org/jira/browse/MATH-1137
> Project: Commons Math
> Issue Type: Bug
> Affects Versions: 3.3
> Reporter: Nigel Goodwin
>
--
This message was sent by Atlassian JIRA
(v6.2#6252)
