You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@commons.apache.org by Gilles Sadowski <gi...@gmail.com> on 2022/05/31 21:52:41 UTC
[Math] GA Design
Hello.
I dug further into the refactoring of the "genetic algorithm" (GA)
functionality of [Math] (in package "o.a.c.math4.legacy.genetics"
currently in the "master" branch of the repository).
This post is a reboot of the discussion thread with Subject:[1]
[Math] Review of "genetic algorithm" module
that is only very slowly converging to what the intended usage is of
the GA implemented in "Commons Math" and the minimal (public)
API for that.
I thus followed up on my initial view[2] of a concise implementation
that avoids the issues which I raised during the previous discussion.
Please have a look at the code committed in branch
feature__MATH-1563__genetic_algorithm
of the repository[3] (in a new "commons-math-ga2" maven module,
for easier comparison with Avijit's proposal).
AFAICT, this design can provide all the functionalities mentioned in the
discussion (a.o. the "adaptive rate"[4]) although it is not complete (and
perhaps contains a few bugs):
* Better names for some classes?
* Should "Population" allow duplicates?
* Port unit tests suite.
* Which genotype representation(s) to support?
* ...
Regards,
Gilles
[1] https://markmail.org/message/2mzdbozc6nwobc37
[2] https://issues.apache.org/jira/browse/MATH-1618
[3] https://gitbox.apache.org/repos/asf?p=commons-math.git;a=shortlog;h=refs/heads/feature__MATH-1563__genetic_algorithm
[4] https://issues.apache.org/jira/browse/MATH-1563
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [Math] GA Design
Posted by Gilles Sadowski <gi...@gmail.com>.
Hello Avijit.
As the OP of the proposal on improving this part of the "Commons
Math" codebase, are you still willing to reach a consensus about
the purpose and design of that functionality?
Regards,
Gilles
Le mar. 31 mai 2022 à 23:52, Gilles Sadowski <gi...@gmail.com> a écrit :
>
> Hello.
>
> I dug further into the refactoring of the "genetic algorithm" (GA)
> functionality of [Math] (in package "o.a.c.math4.legacy.genetics"
> currently in the "master" branch of the repository).
>
> This post is a reboot of the discussion thread with Subject:[1]
> [Math] Review of "genetic algorithm" module
> that is only very slowly converging to what the intended usage is of
> the GA implemented in "Commons Math" and the minimal (public)
> API for that.
>
> I thus followed up on my initial view[2] of a concise implementation
> that avoids the issues which I raised during the previous discussion.
>
> Please have a look at the code committed in branch
> feature__MATH-1563__genetic_algorithm
> of the repository[3] (in a new "commons-math-ga2" maven module,
> for easier comparison with Avijit's proposal).
>
> AFAICT, this design can provide all the functionalities mentioned in the
> discussion (a.o. the "adaptive rate"[4]) although it is not complete (and
> perhaps contains a few bugs):
> * Better names for some classes?
> * Should "Population" allow duplicates?
> * Port unit tests suite.
> * Which genotype representation(s) to support?
> * ...
>
> Regards,
> Gilles
>
> [1] https://markmail.org/message/2mzdbozc6nwobc37
> [2] https://issues.apache.org/jira/browse/MATH-1618
> [3] https://gitbox.apache.org/repos/asf?p=commons-math.git;a=shortlog;h=refs/heads/feature__MATH-1563__genetic_algorithm
> [4] https://issues.apache.org/jira/browse/MATH-1563
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org