[mlpack-git] [mlpack/mlpack] Neural evolution algorithms implementation (CNE, NEAT, HyperNEAT) (#686)

Excalibur notifications at github.com
Sat Jul 16 01:36:56 EDT 2016


> +      }  
> +    }
> +  }
> +
> +  // Crossover two genome to get one genome.
> +  void Crossover(Genome& genome1, Genome& genome2, Genome& childGenome) {
> +    if (CompareGenome(genome1, genome2)) {  // genome1 is better
> +      CrossoverLinkAndNeuron(genome1, genome2, childGenome);
> +    } else {
> +      CrossoverLinkAndNeuron(genome2, genome1, childGenome);
> +    }
> +  }
> +
> +  // Measure two genomes' disjoint (including exceed).
> +  // NOTICE: we can separate into disjoint and exceed. But currently maybe it is enough.
> +  double Disjoint(Genome& genome1, Genome& genome2) {

I plan to change this after NEAT passed Cart Pole problem. As I saw multiple implementations that the parameters for Exceed and Disjoint are the same. Thus, we can keep the current function for now and it won't influence the performance. After that, it can easily be changed.

---
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/mlpack/mlpack/pull/686/files/ff45785b8669f82beb38ef5392f48bcdf5c83f6a#r71061475
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.cc.gatech.edu/pipermail/mlpack-git/attachments/20160715/c3d696e6/attachment.html>


More information about the mlpack-git mailing list