<div dir="ltr">Thanks for your reply<br>I've done some examples using Java libraries as ECJ and JGAP, and was interested in developing my own library to learn gp more in detail.<br>Implementing Push stack based gp seemed very interesting, but if psh is still in active development I think I can help with that.<br>
I will take a closer look to psh.<br>Cheers<br><br><div class="gmail_quote">On Tue, Sep 23, 2008 at 11:03 PM, Lee Spector <span dir="ltr"><<a href="mailto:lspector@hampshire.edu">lspector@hampshire.edu</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"><br>
Pablo,<br>
<br>
Push has been used for a lot of non-mathematical problems, but unfortunately there aren't very many cleaned up examples floating around, and I don't know of an artificial ant example... although it shouldn't be too hard to write.<br>
<br>
In the Lisp version, at <a href="http://hampshire.edu/lspector/push3/" target="_blank">http://hampshire.edu/lspector/push3/</a>, there are a couple of examples including the even n-parity problem (which is Boolean).<br>
<br>
A much more complicated example, using Push but not PushGP per se, is the SwarmEvolve 2.0 system that comes as one of the demos with the breve simulation environment (from <a href="http://www.spiderland.org" target="_blank">www.spiderland.org</a>). This uses many types but it's more of an artificial life application than a GP application.<br>
<br>
Your message is a reminder to me that we are interested in polishing up Jon Klein's new Java implementation called Psh (<a href="http://www.spiderland.org/Psh/" target="_blank">http://www.spiderland.org/Psh/</a>), and a good idea for that would be to code up some other standard problems like artificial ant. But I'm not sure when this will happen.<br>
<br>
On the creation of additional types, I think this may also be easiest in Psh. Maarten Keijzer also has versions in which types can be defined dynamically by Push itself, or built automatically from Java types via reflection, but I don't know if he has made any of these publicly available (I'll cc him).<br>
<br>
-Lee<div><div></div><div class="Wj3C7c"><br>
<br>
<br>
<br>
On Sep 23, 2008, at 12:05 AM, Pablo Ruggia wrote:<br>
<br>
</div></div><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"><div><div></div><div class="Wj3C7c">
Hi<br>
I'm new to the Push language. It seems that it's very good at evolving mathematical calculations, but I can't find much information on other types problems like the the Artificial Ant. Do you know where can I find an example of this kind of problem ?<br>
Also, do you know any implementation of Push that supports the extension of the language to use/create predefined custom types ?<br>
Thanks !<br></div></div>
_______________________________________________<br>
Push mailing list<br>
<a href="mailto:Push@lists.hampshire.edu" target="_blank">Push@lists.hampshire.edu</a><br>
<a href="http://lists.hampshire.edu/mailman/listinfo/push" target="_blank">http://lists.hampshire.edu/mailman/listinfo/push</a><br>
</blockquote>
<br>
--<br>
Lee Spector, Professor of Computer Science<br>
School of Cognitive Science, Hampshire College<br>
893 West Street, Amherst, MA 01002-3359<br>
<a href="mailto:lspector@hampshire.edu" target="_blank">lspector@hampshire.edu</a>, <a href="http://hampshire.edu/lspector/" target="_blank">http://hampshire.edu/lspector/</a><br>
Phone: 413-559-5352, Fax: 413-559-5438<br>
<br>
</blockquote></div><br></div>