OK, I got everything running, wrote a simple bot that's better than the default ones - I had the most trouble figuring out how to get crash callstacks.
My next step is using a high-level grid to control the behavior of my ants (have big cells, keep track of how many ants are in each cell, use this to redirect ants to empty cells and to attract ants to cells with combat. I already did similar things in the past, shouldn't be too complicated).
I may also experiment a bit with coordinating ant movement during fight - if I'm in a cell with a fight, get all my ants to move forward simultaneously or something.
I had thoughts along these lines, but I was thinking that in order to be effective the cells would need to be dynamically sized, and I didn't see an obviously good way to do it (although I since ran across http://en.wikipedia.org/wiki/Local_outlier_factor -- may be of some interest). I'll be curious to know how this turns out for you.
Aichallenge.org has started their third AI contest this year: Ants.
I mentioned this in the open thread, and there was a discussion about possibly making one or more "official" LessWrong teams. D_Alex has offered a motivational prize. If this interests you, please discuss in the comments!