Does anybody have any experience with genetic algorithms, specifically GENESIS?

Hot Topic

determining application path in VS 8, C++ MFC
How to return something from a Void ??
__w64
Resize managed array inside a function
VCBuild and project level settings
Registry editing
rand()'s initial value seems very non-random
VS .Net 2005, how to disable intellisense...
How to add item to a popup menu?
How to list all related functions

Visual C++

build errors in vc++
Error: LNK2019
MessageBox question
Unresolved Token and other link errors when migrating from VC++ 2003 to VC++2005 Express
Shellexecute error
Syntax error declaring CString
PRJ0050
Possible MSVS 2005 beta2 bugs
Visual C++ 2005- class view may lost some classes when those class's header file contains some inline functions.
C2059 error

Does anybody have any experience with genetic algorithms, specifically GENESIS?

Hi,

I have a working model of John Grefenstette's GENESIS genetic algorithm on Microsoft Visual Studio 6.0. My aim now is to get this GA to work as part of a feedback loop with a liquid cystal pulse shaper.

I am trying to write an evaluation function in C++ that will alter the phase and/or amplitude of my LC pulse shaper and the feedback of the system will come from a spectrometer.

I believe that the feedback to my system will be a current generated from the spectrometer but does anyone have any idea how I would code this response into GENESIS to read the signal as the feedback value, and to try and optimize this value

I know this question falls in middle ground as to whether it is science or computer C++ coding, but any help would be much appreciated.

Thanks

Alan Homer




Answer this question

Does anybody have any experience with genetic algorithms, specifically GENESIS?

  • BretUpdegraff

    Thanks.

  • GalenC

    Ok, but you first need some kind of AD converter interface to feed your computer perhaps through your USB or RS-232 port. You need to convert the current that comes from your spectometer into digital values and your genetic algorithm will use those values to correct the shape of your LCPS.

    Also you need some kind of service (driver) that pools your port periodically to collect values (the smaller the pooling period the greater the accuracy you have) and feed your GA classes and methods.

    I guess you will tune your spectometer to analyze below of some threshold and if your LCPS goes beyond, it will feed back a current that is proportional perhaps to some generated armonic noise, amplitude, phase etc.

    Remember that if you want to measure something, you need to have some reference value to compare so in the same way, you need some range of reference your LCPS should be normally and your spectometer must generate a current value (positive or negative) proportional to that shift outside of your reference range, so you can use your algorithm to correct the LCPS to the normal value.

    I hope this help

     



  • Tim erin

    I would try the newsgroups at http://msdn.microsoft.com/newsgroups for such issues. Folks there might be more familiar with similar topics.

    OTP

    Thanks,
    Ayman Shoukry
    VC++ Team


  • Does anybody have any experience with genetic algorithms, specifically GENESIS?
Answer this question