[
https://issues.apache.org/jira/browse/MATH-404?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12896866#action_12896866
]
Luc Maisonobe commented on MATH-404:
------------------------------------
{quote}
Switching between two convergence checking procedures, based on whether a field
is null or not, is at best a temporary workaround, but it is not a good
solution.
{quote}
I agree.
{quote}
Then maybe the scheme needs to be reviewed so that it is general enough to fit.
{quote}
Or LevenbergMarquardtOptimizer needs to be changed and the orthogonality
concept be finally discarded.
{quote}
Allow me to remind what you said: convergence checking is independent from the
optimization algorithm.
But then, in the LM implementation, this doesn't hold...
{quote}
I know, and I am not happy with this. However, I don't want
LevenbergMarquardtOptimizer to be special. It _must_ fit. We can take the
opportunity of a 3.0 major release to fix this problem too, with some
incompatible changes. What would you propose for this ?
> Confusing interface for "LevenbergMarquardtOptimizer"
> -----------------------------------------------------
>
> Key: MATH-404
> URL: https://issues.apache.org/jira/browse/MATH-404
> Project: Commons Math
> Issue Type: Bug
> Affects Versions: 2.1
> Reporter: Gilles
> Fix For: 2.2
>
>
> {{LevenbergMarquardtOptimizer}} inherits from
> {{AbstractLeastSquaresOptimizer}} which in turn implements
> {{DifferentiableMultivariateVectorialOptimizer}}. That interface mandates
> methods for setting and getting a {{VectorialConvergenceChecker}}.
> In v2.1, however, that checker is never used! The convergence check is
> performed using parameters specific to the Levenberg-Marquardt algorithm.
> Such circumvention of the superclass interface is confusing and leads to
> totally unexpected behaviour (such as changing the values of the thresholds
> of the {{VectorialConvergenceChecker}} being ineffective).
> In the development version, the default constructor of
> {{LevenbergMarquardtOptimizer}} sets the the {{VectorialConvergenceChecker}}
> field to "null" and when such is the case, the behaviour is as in v2.1.
> Although it is documented, this is still confusing since it is impossible to
> use {{LevenbergMarquardtOptimizer}} through its
> {{DifferentiableMultivariateVectorialOptimizer}} interface: When using the
> {{VectorialConvergenceChecker}}, one does not know what parameters to use in
> order to reproduce the results obtained with the LM-specific convergence
> check (i.e. how to reproduce the result from v2.1).
> Unless I'm missing something, I think that there should be an LM-specific
> implementation of {{VectorialConvergenceChecker}} that, when given the usual
> relative and absolute thresholds, can perform a check that will give the same
> result as the currently specific check (when the "checker" field is "null").
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
