I didn't know how to create an optimum solution for the ant moves. I did look up constraint optimisation on wikipedia but got very confused, very quickly.
Anyway, I like to solve things my own way, even if it involves reinventing the wheel. For me that was where I had the most fun. I'm afraid I didn't consider comparing my results with the optimum solution so I have no statistics to share with you.
My combat code was an approximation of what "could" happen (it turns out it performs better against the top bots as they tend to avoid the 1:1 swaps that my solution sometimes commits to). The results were good enough.
My exploration code was also good enough. I knew that sometimes 2 ants would go for the same food if they were the same distance away but I wanted the code to stay simple and to not code any special cases.
The issue order code, though not perfect was again good enough once I implemented the "push" concept.
I guess my overall approach was to keep it simple enough for me to understand and therefore simple enough to debug.
Statistics: Posted by Memetix — Mon Dec 19, 2011 10:46 pm