Quantcast

Weighted Roulette Selector - A bug?

classic Classic list List threaded Threaded
9 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Weighted Roulette Selector - A bug?

sethwrkr
This post was updated on .
CONTENTS DELETED
The author has deleted this message.
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Weighted Roulette Selector - A bug?

Paolo Pannarale
I had the same problem which I solved with a fitness lower limit.
BR, Paolo


On Sat, Jan 19, 2013 at 11:35 PM, Seth Meyerson <[hidden email]> wrote:

 

 

From: Seth Meyerson [mailto:[hidden email]]
Sent: Saturday, January 19, 2013 2:56 PM
To: [hidden email]
Subject: Weighted Roulette Selector - A bug?

 

I am implementing a GA program. I am using JGAP.

I am running into an issue with the Weighted Roulette Selector.

The BestChromosome selector works fine.

The Weighted Roulette Selector makes no fitness progress from the 1st generation on. I am dealing with very small fitness values initially. Numbers like .123456 -22, .123456 -09, etc..

Is there any known reason that Roulette will not work in this environment?

Seth


------------------------------------------------------------------------------
Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS,
MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current
with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft
MVPs and experts. ON SALE this month only -- learn more at:
http://p.sf.net/sfu/learnmore_123012
_______________________________________________
jgap-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jgap-users




--
‘[M]y work, which I’ve done for a long time, was not pursued  in order  to gain  the praise  I now enjoy, but chiefly  from a craving after knowledge, which  I notice  resides  in me more  than  in most other men. And wherewithal, whenever I found out anything remarkable, I have thought it my duty to put down my discovery on paper, so that all ingenious people might be informed thereof.’
(van Leeuwenhoek, Letter to the Royal Society, June 12 1716)

“Make it as simple as possible. But not simpler”.
(Albert Einstein)

"The best way to predict the future is to invent it".
(Alan Kay)

___________________________________

Questo documento è indirizzato esclusivamente al destinatario.
Tutte le informazioni ivi contenute, compresi eventuali allegati,
sono soggette a riservatezza secondo i termini del D.Lgs. 196/2003
in materia di "privacy" e ne è proibito l'utilizzo da parte di
altri soggetti. Se avesse ricevuto per errore questo messaggio,
La preghiamo cortesemente di contattare il mittente al più presto
e di cancellare il messaggio subito dopo. Grazie.
----
This document is exclusively intended for the stated addressee.
All information therein, including any attachment, are reserved
as per Italian D.Lgs. 196/2003 about privacy, and cannot be used
by third parties. In case you received this message by mistake,
please inform the sender and delete the message afterward.
Thank you.

------------------------------------------------------------------------------
Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS,
MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current
with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft
MVPs and experts. SALE $99.99 this month only -- learn more at:
http://p.sf.net/sfu/learnmore_122412
_______________________________________________
jgap-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jgap-users
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Weighted Roulette Selector - A bug?

Klaus Meffert-5
In reply to this post by sethwrkr
Seth,
 
I seems as if your assumption regarding the very small fitness values is a good explanation for your problem with WRS.
In generel I would recommend to spread your fitness values (artificially) to a wider range and then apply the selectors.
 
From the past I can say that WRS worked with different use cases.

Best

 


From: Seth Meyerson [mailto:[hidden email]]
Sent: Saturday, January 19, 2013 11:35 PM
To: [hidden email]
Subject: [jgap-users] Weighted Roulette Selector - A bug?

 

 

From: Seth Meyerson [mailto:[hidden email]]
Sent: Saturday, January 19, 2013 2:56 PM
To: [hidden email]
Subject: Weighted Roulette Selector - A bug?

 

I am implementing a GA program. I am using JGAP.

I am running into an issue with the Weighted Roulette Selector.

The BestChromosome selector works fine.

The Weighted Roulette Selector makes no fitness progress from the 1st generation on. I am dealing with very small fitness values initially. Numbers like .123456 -22, .123456 -09, etc..

Is there any known reason that Roulette will not work in this environment?

Seth


------------------------------------------------------------------------------
Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS,
MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current
with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft
MVPs and experts. SALE $99.99 this month only -- learn more at:
http://p.sf.net/sfu/learnmore_122412
_______________________________________________
jgap-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jgap-users
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Weighted Roulette Selector - A bug?

sethwrkr
In reply to this post by sethwrkr
CONTENTS DELETED
The author has deleted this message.
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Weighted Roulette Selector - A bug?

sethwrkr
This post was updated on .
In reply to this post by sethwrkr
CONTENTS DELETED
The author has deleted this message.
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Weighted Roulette Selector - A bug?

Mauricio Noda
Seth,

I looked at WeightedRouletteSelector source code and found out that it has a
precision threshold of 0.000001d. You are using numbers below this thresold,
making the selector assume all fitness values are equal and behave like a
random selector.

You need to increase the difference between fitness values somehow.

-----Original Message-----
From: sethwrkr
Sent: Wednesday, January 23, 2013 11:37 PM
To: [hidden email]
Subject: Re: [jgap-users] Weighted Roulette Selector - A bug?


Just weird.  I used tournament with a tourney size of 4 and 1.00 as the
winning percentage.  No hits on optimal solutions.  and Just to reiterate, 0
hits with roulette.

with Bestchromosome I get 95% of 10,000 different symptom sets.




sethwrkr wrote:

>
>
>
>
>
> From: Seth Meyerson [mailto:[hidden email]]
> Sent: Saturday, January 19, 2013 2:56 PM
> To: [hidden email]
> Subject: Weighted Roulette Selector - A bug?
>
>
>
> I am implementing a GA program. I am using JGAP.
>
> I am running into an issue with the Weighted Roulette Selector.
>
> The BestChromosome selector works fine.
>
> The Weighted Roulette Selector makes no fitness progress from the 1st
> generation on. I am dealing with very small fitness values initially.
> Numbers like .123456 -22, .123456 -09, etc..
>
> Is there any known reason that Roulette will not work in this environment?
>
> Seth
>
>
> ------------------------------------------------------------------------------
> Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS,
> MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current
> with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft
> MVPs and experts. ON SALE this month only -- learn more at:
> http://p.sf.net/sfu/learnmore_123012
> _______________________________________________
> jgap-users mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/jgap-users
>
>

--
View this message in context:
http://old.nabble.com/Weighted-Roulette-Selector---A-bug--tp34923528p34938175.html
Sent from the jgap-users mailing list archive at Nabble.com.


------------------------------------------------------------------------------
Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS,
MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current
with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft
MVPs and experts. ON SALE this month only -- learn more at:
http://p.sf.net/sfu/learnnow-d2d
_______________________________________________
jgap-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jgap-users 


------------------------------------------------------------------------------
Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS,
MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current
with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft
MVPs and experts. ON SALE this month only -- learn more at:
http://p.sf.net/sfu/learnnow-d2d
_______________________________________________
jgap-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jgap-users
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Weighted Roulette Selector - A bug?

sethwrkr
In reply to this post by Klaus Meffert-5
CONTENTS DELETED
The author has deleted this message.
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Weighted Roulette Selector - A bug?

sethwrkr
CONTENTS DELETED
The author has deleted this message.
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Weighted Roulette Selector - A bug?

Klaus Meffert-5
Seth,

yes, the problem seems to be related to the small fitness values and how WRS
handles them. I sent you an updated version of the WeightedRouletteSelector
class last week.
I also checked in the new version to CVS.

Best

Klaus
www.klaus-meffert.com

 

> -----Original Message-----
> From: sethwrkr [mailto:[hidden email]]
> Sent: Thursday, January 31, 2013 3:36 PM
> To: [hidden email]
> Subject: Re: [jgap-users] Weighted Roulette Selector - A bug?
>
>
> Kalus,
>
> What about the specific precision issue noted by Mauricio
> Noda?  Is there a difference in precision between best
> chromosome (ranked selection) and weighted roulette's ability
> to distinguish between very  small and close
> fitness values.  
>
> His assertion matches my experience.  Best chromosome works well.
> Tournament and Roulette do not progress towards a solution
> with the fitness function I implemented.
>
> Seth
>
>
> sethwrkr wrote:
> >
> > Klaus,
> >
> > Is Mauricio Noda correct on the precision issue?
> >
> > if so is there any way to improve the precision of weighted
> roulette
> > selector to match best chromosome?
> >
> > Seth
> >  
> > I seems as if your assumption regarding the very small
> fitness values
> > is a good explanation for your problem with WRS.
> > In generel I would recommend to spread your fitness values
> > (artificially) to a wider range and then apply the selectors.
> >  
> >>From the past I can say that WRS worked with different use cases.
> >
> > Best
> >
> > Klaus
> > www.klaus-meffert.com <http://www.klaus-meffert.com/>
> >
> >  
> >
> >
> >   _____
> >
> > From: Seth Meyerson [mailto:[hidden email]]
> > Sent: Saturday, January 19, 2013 11:35 PM
> > To: [hidden email]
> > Subject: [jgap-users] Weighted Roulette Selector - A bug?
> >
> >
> >
> >  
> >
> >  
> >
> > From: Seth Meyerson [mailto:[hidden email]]
> > Sent: Saturday, January 19, 2013 2:56 PM
> > To: [hidden email]
> > Subject: Weighted Roulette Selector - A bug?
> >
> >  
> >
> > I am implementing a GA program. I am using JGAP.
> >
> > I am running into an issue with the Weighted Roulette Selector.
> >
> > The BestChromosome selector works fine.
> >
> > The Weighted Roulette Selector makes no fitness progress
> from the 1st
> > generation on. I am dealing with very small fitness values
> initially.
> > Numbers like .123456 -22, .123456 -09, etc..
> >
> > Is there any known reason that Roulette will not work in
> this environment?
> >
> > Seth
> >
> >
> >
> ----------------------------------------------------------------------
> > -------- Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012,
> > HTML5, CSS, MVC, Windows 8 Apps, JavaScript and much more.
> Keep your
> > skills current with LearnDevNow - 3,200 step-by-step video
> tutorials
> > by Microsoft MVPs and experts. SALE $99.99 this month only -- learn
> > more at:
> > http://p.sf.net/sfu/learnmore_122412
> > _______________________________________________
> > jgap-users mailing list
> > [hidden email]
> > https://lists.sourceforge.net/lists/listinfo/jgap-users
> >
> >
>
>
>
> --
> View this message in context:
> http://old.nabble.com/Weighted-Roulette-Selector---A-bug--tp34
923528p34967482.html

> Sent from the jgap-users mailing list archive at Nabble.com.
>
>
> --------------------------------------------------------------
> ----------------
> Everyone hates slow websites. So do we.
> Make your web apps faster with AppDynamics
> Download AppDynamics Lite for free today:
> http://p.sf.net/sfu/appdyn_d2d_jan
> _______________________________________________
> jgap-users mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/jgap-users


------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_d2d_jan
_______________________________________________
jgap-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jgap-users
Loading...