Quantcast

problem definition

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

problem definition

Nikolaos Papadakis-2
Hi All,

I have the following problem, I hope you can help me.

I have a repository of registered members. Each member owns a set of attributes e.g. age, geographical location, sex, level of expertise, etc.

I want to collect a certain number of surveys (lets say X) by inviting X members.
At the end of the process I want to have X surveys with a particular segmentation based on the attributes (e.g. I need 30% to be female, 40% from the south part of the country and 30% of a certain level of expertise).
What I the best example to base on this problem ? I don't know how to begin...

Thank you

Nikos
------------------------------------------------------------------------------
Colocation vs. Managed Hosting
A question and answer guide to determining the best fit
for your organization - today and in the future.
http://p.sf.net/sfu/internap-sfd2d
_______________________________________________
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: problem definition

Klaus Meffert-5
Hello Nikos,
 
I am not if I understand your seeting correctly: Isn't it possible to do the collection (30% femal, 40% from south...) by just examining all surveys?
What should be the genetic algorithms's role here?
 
Best
 


From: Nikolaos Papadakis [mailto:[hidden email]]
Sent: Tuesday, March 08, 2011 4:06 PM
To: [hidden email]
Subject: [jgap-users] problem definition

Hi All,

I have the following problem, I hope you can help me.

I have a repository of registered members. Each member owns a set of attributes e.g. age, geographical location, sex, level of expertise, etc.

I want to collect a certain number of surveys (lets say X) by inviting X members.
At the end of the process I want to have X surveys with a particular segmentation based on the attributes (e.g. I need 30% to be female, 40% from the south part of the country and 30% of a certain level of expertise).
What I the best example to base on this problem ? I don't know how to begin...

Thank you

Nikos

------------------------------------------------------------------------------
Colocation vs. Managed Hosting
A question and answer guide to determining the best fit
for your organization - today and in the future.
http://p.sf.net/sfu/internap-sfd2d
_______________________________________________
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: problem definition

Nikolaos Papadakis-2
Thank you for your response.
You are right it is just an “arithmetic” problem. I managed to solve it with a more classical rationale :-)
However now I have a different problem that seams to be related with the JGAP (?):

I have a set of papers each of one belonging in one or more predefined categories.
I also have a (larger) set of reviewers each of one claiming one or more expertises (from the same papers' category - list let denote it as category vector).
For instance we have a paper p1 belonging to A, B, C, and D category and we have a reviewer r1 claiming to be expert in categories B and D from the same category vector.
I tried to solve this by simply performing a clustering of the reviewers so I have a number of clusters each of them containing a number of reviewers with similar expertises.
Then by processing each paper one by one I measure its distance form each cluster of reviewers and then I assign it to the reviewer that is found closer (in terms of the defined distance measure) the reviewer then is not taken into consideration anymore (this is for simplicity in reality each reviewer should accept a predefined number of papers – lets say 3 papers and then he should be “completed”).
It turns out that at the end there may be papers that can't be assigned because they have nothing in common with the remaining reviewers. However if I change the order that I process the papers (e.g order by name not by id) then the problem may converge to a different solution that is more “acceptable” (that means it can converge to a solution with a few or zero non assigned papers – orphans).
I tried to find something more fine in  the scientific literature and I found the “Stable Marriage Problem” (http://en.wikipedia.org/wiki/Stable_marriage_problem ) but it requires the same number of reviewers and papers and this is not my case since only in extreme cases these two sets have the same cardinality. In general I have a larger number of reviewers that will accept at most 3 papers and each paper that should be assigned at most to 5 reviewers.
Do you think this can be accomplished by using the JGAP ?

Thank you in advance,

Nikos Papadakis
   


On Thu, Mar 17, 2011 at 11:01 AM, Klaus Meffert <[hidden email]> wrote:
Hello Nikos,
 
I am not if I understand your seeting correctly: Isn't it possible to do the collection (30% femal, 40% from south...) by just examining all surveys?
What should be the genetic algorithms's role here?
 
Best
 


From: Nikolaos Papadakis [mailto:[hidden email]]
Sent: Tuesday, March 08, 2011 4:06 PM
To: [hidden email]
Subject: [jgap-users] problem definition

Hi All,

I have the following problem, I hope you can help me.

I have a repository of registered members. Each member owns a set of attributes e.g. age, geographical location, sex, level of expertise, etc.

I want to collect a certain number of surveys (lets say X) by inviting X members.
At the end of the process I want to have X surveys with a particular segmentation based on the attributes (e.g. I need 30% to be female, 40% from the south part of the country and 30% of a certain level of expertise).
What I the best example to base on this problem ? I don't know how to begin...

Thank you

Nikos


------------------------------------------------------------------------------
Benefiting from Server Virtualization: Beyond Initial Workload
Consolidation -- Increasing the use of server virtualization is a top
priority.Virtualization can reduce costs, simplify management, and improve
application availability and disaster protection. Learn more about boosting
the value of server virtualization. http://p.sf.net/sfu/vmware-sfdev2dev
_______________________________________________
jgap-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/jgap-users
Loading...