[mlpack] Particle swarm optimization for GSOC-19

[Marcus Edel]

I like the idea, it might be helpful to checkout the existing constrained
interface: http://www.ensmallen.org/docs.html#constrained-functions as well.

> On 22. Mar 2019, at 03:50, sangyunxin <sangyunxin at gmail.com> wrote:
> 
> Hi Marcus, 
> Thank you very much for your explanation. It was great help. 
> For this project,  I think we can use penalty method <https://en.wikipedia.org/wiki/Penalty_method> or some improved algorithm based on this. The essence of this method lies in replacing a constrained problem by a series of unconstrained problems. By this means, we may come up with a working optimizer which supports both.
> Please let me know what you think.
> Thank you for your time and consideration.
> Sent from YoMail for Gmail <http://www.yomail.com/?utm_source=signature>
> 发件人： Marcus Edel <mailto:marcus.edel at fu-berlin.de>
> 发送时间： 2019-03-21 06:25:17
> 收件人：  sangyunxin <mailto:sangyunxin at gmail.com>
> 抄送：  mlpack <mailto:mlpack at lists.mlpack.org>
> 主题： Re: [mlpack] Particle swarm optimization for GSOC-19
> Hello Sang,
> 
> welcome, you are right ideally we can come up with an implementation that
> supports both, but if we can't e.g. if we are going to implement a PSO variant
> that only works on one there is no need to merge both approaches.
> 
> I hope anyything I said makes sense, let me know if I should clarify anything.
> 
> Thanks,
> Marcus
> 
>> On 19. Mar 2019, at 03:07, sangyunxin <sangyunxin at gmail.com <mailto:sangyunxin at gmail.com>> wrote:
>> 
>> Dear all,
>> My name is Sang Yunxin, and I am a senior student from Whuhan University, China. I want to participate in GSoC 2019 and take part in establishing "Particle swarm optimization". 
>> According to my understanding of this idea and previous discussions about this project, I will work to provide a PSO optimizer which works for both unconstrained and constrained problems to complement mlpack optimizer API —— ensmallen. Do I have a correct understanding for this project? 
>> Thank you, and I look forward to hearing from you.
>> _______________________________________________
>> mlpack mailing list
>> mlpack at lists.mlpack.org <mailto:mlpack at lists.mlpack.org>
>> http://knife.lugatgt.org/cgi-bin/mailman/listinfo/mlpack <http://knife.lugatgt.org/cgi-bin/mailman/listinfo/mlpack>
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://knife.lugatgt.org/pipermail/mlpack/attachments/20190324/8ec49925/attachment.html>