Mailing-List: contact issues-help@commons.apache.org; run by ezmlm
Precedence: bulk
Reply-To: issues@commons.apache.org
Date: Sun, 9 Dec 2012 17:13:20 +0000 (UTC)
From: "Gilles (JIRA)" <jira@apache.org>
To: issues@commons.apache.org
Message-ID: <JIRA.12618547.1354531268947.9508.1355073200751@arcas>
In-Reply-To: <JIRA.12618547.1354531268947@arcas>
References: <JIRA.12618547.1354531268947@arcas>
Subject: [jira] [Commented] (MATH-911) Remove usage of "DerivativeStructure"
 in the "optimization" package
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/MATH-911?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13527528#comment-13527528 ] 

Gilles commented on MATH-911:
-----------------------------

Following the changes proposed in MATH-874, we could resolve this as "Won't fix": the current code is backward-compatible, and *all* classes in
* o.a.c.m.optimization
* o.a.c.m.optimization.direct
* o.a.c.m.optimization.general
* o.a.c.m.optimization.univariate

are going to be deprecated (and replaced by those in package "o.a.c.m.optim", see MATH-874); thus reverting codes that will deleted afterwards seems to be unnecessary work.

There is still the issue with the package
* o.a.c.m.optimization.fitting

Classes in that package also use "DerivativeStructure".
Is it necessary or required? If so, adapters will probably be needed when calls are made to the new classes in package "o.a.c.m.optim" where all references to "DerivativeStructure" have been removed.
If not necessary, then those classes have yet to be reverted to not use "DerivativeStructure".

                
> Remove usage of "DerivativeStructure" in the "optimization" package
> -------------------------------------------------------------------
>
>                 Key: MATH-911
>                 URL: https://issues.apache.org/jira/browse/MATH-911
>             Project: Commons Math
>          Issue Type: Task
>            Reporter: Gilles
>            Priority: Minor
>             Fix For: 3.1
>
>
> The "DerivativeStructure" class is not the proper interface towards user applications that deal with optimization problems.
> Recent changes implies copying between e.g. the user's representation of a Jacobian and the "DerviativeStructure" representation of the same data. And, actually, the current implementation of "AbstractLeastSquaresOptimizer" copies back the "DerviativeStructure" representation into a "double[][]" for internal use.
> [~kberlin] pointed out the potential storage inefficiencies in large-scale problems.

--
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