Loading ...
Sorry, an error occurred while loading the content.
 

Re: {Disarmed} RE: [GP] A question about Genetic Programming

Expand Messages
  • okerol@itu.edu.tr
    Hi Hazem, You may also try our Big-Bang big-Crunh method for finding the optimum parameter set of a given function. This method converges to global optimum
    Message 1 of 6 , Dec 30, 2007
      Hi Hazem,
      You may also try our Big-Bang big-Crunh method for finding the optimum
      parameter
      set of a given function.
      This method converges to global optimum 1000+ faster than Genetic Algorithms.
      You may download the MATLAB codes at my web site at:

      www3.itu.edu.tr/~okerol

      The principle is: Like GA, the algorithm chooses randomly a number of possible
      choices (chromosomes in GA) calculates their cost function then chooses the
      minimum of them. Then by using normal probability distribution generates new
      candidates around that point.

      Hi Thomas,
      You may also use that method for Genetic Programming "after" obtaining a
      structure, to find the constants that may appear in that S-expression. That
      speeds the Genetic search procedure.

      Kind Regards,
      Osman




      Quoting Thomas Weise <tweise@...>:

      > Hello Hazem.
      >
      >
      >
      > Hm, if you want to find minima or maxima of a _given_ function,
      >
      > Genetic Programming may not be the approach of choice.
      >
      > Here, other methods like Evolution Strategy/Differential Evolution
      >
      > or Particle Swarm Optimization may be better.
      >
      > Example: f(x)=x*x-3x -> Find the minimum of this function.
      >
      > Answer (by PSO): Minimum f(x)=-2.25 is at x=1.5
      >
      >
      >
      > With Genetic Programming we normally breed tree-shaped structure.
      >
      > Since mathematical functions are in principle hierarchies of
      >
      > mathematical expressions which can in turn be expressed as trees,
      >
      > GP is a good means to _find_ functions that fit to given sample
      >
      > data.
      >
      > Example: A set of values (x,f(x)) is given like {(0,0); (1,-2); (2,-2).
      >
      > -> Find the function f(x) that fits to these data samples.
      >
      > -> Answer (by GP): f(x)=x*x-3x
      >
      >
      >
      > I have tried to compile a summary on different evolutionary
      >
      > algorithms. Maybe it can be helpful for you. You can find it at
      >
      > http://www.it-weise.de/projects/book.pdf
      >
      >
      >
      > Kind regards,
      >
      > Thomas.
      >
      >
      >
      > From: genetic_programming@yahoogroups.com
      > [mailto:genetic_programming@yahoogroups.com] On Behalf Of Hazem Saleh
      > Sent: Tuesday, December 25, 2007 5:12 AM
      > To: genetic_programming@yahoogroups.com
      > Subject: [GP] A question about Genetic Programming
      >
      >
      >
      > Hi Guys;
      > Iam a beginner at the genetic programming.
      > I have a question which may be a stupid one.
      > Can I use genetic programming for function (with several independent
      > variables) optimization like genetic algorithms?
      > If the answer is yes; Is there a specific tutorial that can help me to
      > realize how this can be done.
      > Iam sorry again if my question is stupid; Iam a beginner.
      > Thanks all!
      >
      >
      >
      >
      >
      > [Non-text portions of this message have been removed]
      >
      >



      ----------------------------------------------------------------
      This message was sent using IMP, the Internet Messaging Program.
    • aichour malek
      Hello, i s a good idea to use GP for independant variable but i haven t try it but i think that it can be possible. if you find any think about this you can
      Message 2 of 6 , Dec 31, 2007
        Hello,
        i's a good idea to use GP for independant variable but i haven't try
        it but i think that it can be possible. if you find any think about
        this you can write me perhaps i can help in th efitness function.
        thank's

        ----------------------------------------------------------------
        This message was sent using IMP, the Internet Messaging Program.
      • David vun Kannon
        Does anyone know of an online source for this article? I have not been able to track it down. Thanks, David vun Kannon @inproceedings{DBLP:conf/icga/Paredis95,
        Message 3 of 6 , Jan 9, 2008
          Does anyone know of an online source for this article? I have not been able
          to track it down.

          Thanks,
          David vun Kannon


          @inproceedings{DBLP:conf/icga/Paredis95,
          author = {Jan Paredis},
          title = {The Symbiotic Evolution of Solutions and Their
          Representations},
          booktitle = {ICGA},
          year = {1995},
          pages = {359-365},
          crossref = {DBLP:conf/icga/1995},
          bibsource = {DBLP, http://dblp.uni-trier.de}
          }
        Your message has been successfully submitted and would be delivered to recipients shortly.