It appears that single step mutation is fully operational. There do appear to be some bugs here and there, but I should be able to find them eventually.
Single step mutation is where we only have one batch of mutants, so rules aren’t continually mutated. Further steps could result from mutating extra rules every CE update iteration, where we sample the distribution N times (where N is equal to the number of rules in the slot).
A further problem is re-covering, where it isn’t necessary. This could be fixed by appending the general rules (in sampled order at the end of every policy, as default go-to rules should the more specific ones fail.
This entry was posted on Monday, March 8th, 2010 at 11:58 am and is filed under PhD Project. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.