Any human-like behavior must take into account making mistakes. I’m thinking about this today and how to make a self-correcting system.
Written by MM. Posted at 5:25 am on February 5th, 2010
8 comments.
Post a comment.
Hmm, adding all of these pieces to the puzzle is going to have a deep impact on game performance.
Maybe having each computer that is in the game take part of the calculation (for example each computer calculating behaviour of 1 bot) would be a solution for that? And this brings us back to the peer-to-peer game, where nobody is THE server. Any thoughts on that lately?
@Tenshi
How can you know it’s going to have a deep impact on game performance?
Easy. More systems = More calculations = need more performance. Bam.
@Tenshi that’s an oversimplification to the point of being meaningless. Adaptive decision making is almost certainly NOT something the game does during every frame.
It would be ideal if it did it every frame cause that’s what humans do. But right now it does calculated decisions only when needed.
MM, I would like to hear that you’ll be cautious and you won’t make AI that will turn into Skynet and doom humanity in near future, so we won’t have to live in real life Link-Dead post apocalyptic nightmare 😉
Rook_PL: I’m afraid that if the AI decides that the best way to kill you will be launch a nuke at your city where you live then I have no control over that.
Благодарю за пост. Очень вовремя.