Hi Uday,

Thanks for the reply.

I will check that out, although I’m really keen to find an implementation of Poli and McPhee’s method — it seems like the most principled method in the literature, in that it is theoretically-derived.



On 13 Apr 2016, at 16:00, Uday kamath <[log in to unmask]> wrote:

Look at the https://github.com/eclab/ecj/blob/master/ecj/ec/gp/breed/SizeFairCrossoverPipeline.java

Found it fairly doing well in controlling the bloat.


On Wed, Apr 13, 2016 at 10:37 AM, White, David <[log in to unmask]> wrote:
Hi Everyone,

In 2008, Poli and McPhee published “Parsimony Pressure Made Easy: Solving the Problem of Bloat in GP.”

In the paper, they outline a fairly straightforward way of controlling mean program size in a GP population.

Has anyone implemented this in ECJ, please? I’d like to use it and don’t want to repeat the effort if it’s already been done.