 
| Mime-Version: |
1.0 (Apple Message framework v746.2) |
| Sender: |
|
| Date: |
Fri, 30 Dec 2005 12:01:15 -0800 |
| Reply-To: |
|
| Subject: |
|
| From: |
|
| Content-Transfer-Encoding: |
7bit |
| In-Reply-To: |
|
| Content-Type: |
text/plain; charset=US-ASCII; format=flowed |
| Comments: |
|
| Parts/Attachments: |
|
|
Perfect -- thanks!
On Dec 30, 2005, at 11:37 AM, Liviu Panait wrote:
> I think the easiest approach is to implement your own statistics
> (maybe extend SimpleStatistics).
>
>> 1) How do I save the best-of-generation individual to a file each
>> generation (or every N generations)? My individuals are too large to
>> print, and I'd like to do some other things with them after I evolve
>> them.
>
> In your new statistics, you just print that individual wherever you
> want (screen, file, etc).
>
>> 2) As mentioned, I'm coevolving Go players. To see that I'm making
>> progress, I'd like to test the best-of-generation individual against
>> a random player and against a known good player every so many
>> generations. Is there any easy way to do this?
>
> Again, you can do that in your custom statistics. The
> postEvaluationStatistics method looks perfect for both tasks.
>
> Liviu.
Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/
|
|
|
