3 ms·
1) "then there probably exists a very simple optimal strategy, and it's more a problem of mathematics than AI." I don't see how existence of the optimal strate
by xyzzyz 16y ago
1) "then there probably exists a very simple optimal strategy, and it's more a problem of mathematics than AI."
I don't see how existence of the optimal strategy entails that this is not the problem of AI. After all, all finite state deterministic games can be essentially solved by traversing the game tree, searching for the best strategy. The thing is that in most games, this approach is not acceptable because of time and space issues. That's why we call for AI - to find a strategy which may not be optimal, but is "good enough" and does not take much resources to find.
Besides, I do not understand why do you think that simple optimal strategy exists. The game of Go has equally simple rules, finding good strategy in it is not easy though.
2) The rules are not clear for me either. Last night I wrote a simple engine for planet wars [1], and I had to make some important decisions about things which the game description did not cover - for instance, what is the order of events in game? Is the number of ships on players' planets increased before or after the fleets arrive to them?
[1] - http://bitbucket.org/dodek/gaclon/src http://bitbucket.org/dodek/gaclon/src