From fr.teodoro at gmail.com Wed Jun 11 14:30:25 2008 From: fr.teodoro at gmail.com (=?ISO-8859-1?Q?F=E1bio_Roberto_Teodoro?=) Date: Wed, 11 Jun 2008 15:30:25 -0300 Subject: [galib] GARealGenome with SwapMutator doubt Message-ID: <485019C1.5010700@gmail.com> If I set diferent bounds for genes and use the SwapMutator. What happens? Looking at the code of GAlib I was thinking tha it could result in invalid individuals, but in my tests it doesn't happen. Is SwapMutator doing nothing? TIA From sandipkarale at gmail.com Thu Jun 19 06:07:06 2008 From: sandipkarale at gmail.com (Sandip Karale) Date: Thu, 19 Jun 2008 15:37:06 +0530 Subject: [galib] Reproducing the results > GARealGenome Message-ID: Hello all, I'm using real genome. Is there any other way reproduce the result. I have tried using seed but it was not helping. Thanks, Sandi -- Rodney Dangerfield - "I haven't spoken to my wife in years. I didn't want to interrupt her." -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mailman.mit.edu/pipermail/galib/attachments/20080619/cdc285b4/attachment.htm From bigblue_welly86 at yahoo.com Sat Jun 21 01:53:28 2008 From: bigblue_welly86 at yahoo.com (Wellya Septin) Date: Fri, 20 Jun 2008 22:53:28 -0700 (PDT) Subject: [galib] (no subject) In-Reply-To: <485019C1.5010700@gmail.com> Message-ID: <687013.64185.qm@web52803.mail.re2.yahoo.com> -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mailman.mit.edu/pipermail/galib/attachments/20080620/22be42b6/attachment.htm From fr.teodoro at gmail.com Sat Jun 21 13:24:22 2008 From: fr.teodoro at gmail.com (=?ISO-8859-1?Q?F=E1bio_Roberto_Teodoro?=) Date: Sat, 21 Jun 2008 14:24:22 -0300 Subject: [galib] About the objetive function Message-ID: <485D3946.9050902@gmail.com> I have some doubts about the objective function. It means the raw fitness, the value of the function that I am optimizing or something else? I think the it mean the second option, but not sure. And why it gives me error messages when the objective function returns negative numbers? If someone could clarify this points... TIA From kskk02 at gmail.com Sat Jun 21 16:44:31 2008 From: kskk02 at gmail.com (Sri Kanajan) Date: Sat, 21 Jun 2008 13:44:31 -0700 Subject: [galib] About the objetive function In-Reply-To: <485D3946.9050902@gmail.com> References: <485D3946.9050902@gmail.com> Message-ID: the objective function represents your optimization goal. In mathematical programming, this is explicitly defined as a maximization or minimization of a specific function that tends to be dependent on your decision variables. In a GA formulation, the objective function is reflected by the fitness funtion. Essentially, this function is evaluated for every generation of alleles that is generated. The fitness function is a function of all the alleles for a specific individual. Essentially this score represents how good that specific solution is. As the genetic algorithm progresses and generates new individuals, individuals with a higher score are kept and used to generate mutated individuals that potentially could have a better score. Hope this helps. sri On Sat, Jun 21, 2008 at 10:24 AM, F?bio Roberto Teodoro < fr.teodoro at gmail.com> wrote: > I have some doubts about the objective function. It means the raw fitness, > the > value of the function that I am optimizing or something else? I think the > it > mean the second option, but not sure. > > And why it gives me error messages when the objective function returns > negative > numbers? > > If someone could clarify this points... > > TIA > _______________________________________________ > galib mailing list > galib at mit.edu > http://mailman.mit.edu/mailman/listinfo/galib > -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mailman.mit.edu/pipermail/galib/attachments/20080621/2474eb22/attachment.htm From fr.teodoro at gmail.com Fri Jun 27 01:15:54 2008 From: fr.teodoro at gmail.com (=?ISO-8859-1?Q?F=E1bio_Roberto_Teodoro?=) Date: Fri, 27 Jun 2008 02:15:54 -0300 Subject: [galib] TerminateUponConvergence problem Message-ID: <4864778A.1030100@gmail.com> When the AG converges to a score of 0 the terminator TerminateUponConvergence never returns gaTrue. Is this a bug in GAlib? Thanks for any help From eyaler at yahoo.com Sat Jun 28 07:57:48 2008 From: eyaler at yahoo.com (Eyal) Date: Sat, 28 Jun 2008 04:57:48 -0700 (PDT) Subject: [galib] TerminateUponConvergence problem In-Reply-To: <4864778A.1030100@gmail.com> Message-ID: <855123.41091.qm@web31403.mail.mud.yahoo.com> yes, this is in fact the case you can add something like: if (ga.population().best().score()==0.0f&&ga.generation()>=ga.nConvergence()) break; in your code to circumvent the bug --- On Fri, 6/27/08, F?bio Roberto Teodoro wrote: > From: F?bio Roberto Teodoro > Subject: [galib] TerminateUponConvergence problem > To: galib at mit.edu > Date: Friday, June 27, 2008, 8:15 AM > When the AG converges to a score of 0 the terminator > TerminateUponConvergence > never returns gaTrue. Is this a bug in GAlib? > > Thanks for any help > _______________________________________________ > galib mailing list > galib at mit.edu > http://mailman.mit.edu/mailman/listinfo/galib