Hello! I'm using OpenBeagle 3.0.3 and applying it to symbolic regression. It's just awesome software!, however few things are not clear to me. I have a quad...
Hi Sergiy, Fitness evaluation in Open BEAGLE is not designed for parallel execution on multi-core computers. I agree that it would be very interesting for ...
Hello Christian, Thank you very much for such a detailed response. I didn't get it working yet, but at least now there are problems with my code. However I...
Hi Sergiy, Standard GP tree interpretation modifies only values in the Context. Therefore, as far as you have one Context per thread, parallel fitness ...
Hi, In my program where I co-evolve GP populations using my own threading code, as discussed in July and August I've been stress testing things and get perfect...
Hi Stig, The implementation of ADFs in Open BEAGLE is inherently not thread-safe. The problem is with the GP::Argument primitive, which store locally the value...
Hi again, Just one precision, following message from Stig. Argument terminals in GP programs make, for example the X values in symbolic expressions, are ...
... I'm afraid that even after changing it to just-in-time mode I still get the same crashes :( I don't actually do concurrent fitness evaluation as I alsways...
That's what I meant. I do set several variables (GP::EvaluationOP::setValue) for individual and was suspicious about that. I guess I will follow co-ev path...
... It was just coincidence - it crashed a little while later :( The reason I asked about how cloning of individuals with ADFs works, is that if references...
Hi Stig, ... I don't think there should be problems with the ADFs and cloning data in general. This is thread-safe. But as I explained to Sergyi in another...
Hello! Is there a simple way to create individual with primitive tree, which would represent expression: (X + Y) * 5? X and Y are variables which can be set...
Hi Sergiy, GP trees are simply vector of nodes organized in prefix order. http://en.wikipedia.org/wiki/Polish_notation First, you must have a primitive set...
Thank you very much Christian! Finally I got it to work. Here is the code in case someone is interested: GP::PrimitiveSet::Handle lSet = new GP::PrimitiveSet; ...
Hi Sergiy, Thanks for the information. More below. ... This is the index of the PrimitiveSet to use in the PrimitiveSuperSet. The PrimitiveSuperSet is simply a...
Hello all, Hopefully this will be usefully for someone. Few weeks ago I realised that there are no obvious bottlenecks in OpenBeagle and the only way to...
Hi Pascal, For reading the best individual from the milestone file, you need to craft some code to extract it from the Vivarium's hall-of-fame. For that ...
Hi, The most convenient way to read an individual from a file is to use method Beagle::Individual::readFromFile(std::string inFileName, Beagle& ioSystem). The...
Hi, Ah, sorry, I forgot the fact that you need to set the genotype allocator correctly set-up in order to read the individual. GP::Tree::Alloc::Handle hTree =...
Hi Pascal, Look at the configuration file 'parity-adfs.conf'. It shows how ADFs can be set with the framework. In general, no coding is necessary, only good ...
Hello, I am new in Open Beagle. I seems to be a great, complex framework! I am trying now to play with the examples and learn as much as possible from the...
Hi, ... That's mostly it. A reference to the yahoogroups list is given on the project's Web page. http://beagle.gel.ulaval.ca/#mailinglist There is also a...
Hi, I noticed that GA examples uses GA::FloatVector or GA::IntegerVector as a solution vector. Is it possible to create mixed type vector? For instance I want...
Hi, That's possible, but then you need to create a custom genotype. Open BEAGLE is made for that. Many operators can be used as is as they don't depend on the...
I'm trying to implement extended statistics that would measure the diversity of the population on genotype/phenotype/fitness level. I have found this two not...
Paweł Widera
plw@...
Feb 2, 2009 2:20 am
1313
Hi all, I'm just wondering if Open Beagle supports survival ratio evolutionary scheme. For example, the populations size is 100, and only 30 out of the whole...
Hi Pawel, ... I added these methods as something that can be reused in some particular circumstances, as so need at least to be defined in the "mother"...
Hi Linge, This approach is not specifically supported by Open BEAGLE. So short answer is yes, you will need to implement this yourself. But, do you have any...