Is there a good book for AI implementation that isn't centered around FPS or RTS games?
I'm looking for something that covers the more basic "old-fashioned" problems like how to actually write a chess/checkers AI or GO or basically any board game.
Obviously I will need to tailor specifics to what I'm trying to achieve, but I just don't know how to write a good "brute force" algorithm, or what a negamax actually looks like, etc.
Thanks in advance...
Book Recommendation ?
Barring that, there is a ridiculous number of articles on the web that show a lot of that already.
Author
Care to point any out? One of my problems is that I know so little about this topic that I can't even phrase the right questions at this point.
My particular problem is a fairly simple board game (with very few rules) for up to 7 players. I need to brute force a couple of moves ahead and pick a best move. That's it really.
My particular problem is a fairly simple board game (with very few rules) for up to 7 players. I need to brute force a couple of moves ahead and pick a best move. That's it really.
Seriously?
Negamax pseudocode per Wikipedia
Tons of GameDev threads on negamax
As for "brute force", there is no such thing in a generalized sense. All it means is that you do something horribly inefficient to try and check every possible situation. (Often you see a lot of nested loops iterating through all possibilities.)
Negamax pseudocode per Wikipedia
Tons of GameDev threads on negamax
As for "brute force", there is no such thing in a generalized sense. All it means is that you do something horribly inefficient to try and check every possible situation. (Often you see a lot of nested loops iterating through all possibilities.)
Author
Thanks for the links.
And yes, absolutely serious. I've been in pro games dev for 25 years, but there's always been AI specialists in-house, so I'm a long way behind the curve on this one.
The tanks game you might remember me asking about is my only other forray into AI, and both have come in the last few months so I guess I need to start boning up rapido, hence the book recommendations and stuff.
I do understand enough from general osmosis that negamax isn't suitable for seven players though :(
And yes, absolutely serious. I've been in pro games dev for 25 years, but there's always been AI specialists in-house, so I'm a long way behind the curve on this one.
The tanks game you might remember me asking about is my only other forray into AI, and both have come in the last few months so I guess I need to start boning up rapido, hence the book recommendations and stuff.
I do understand enough from general osmosis that negamax isn't suitable for seven players though :(
My point was in reference to the fact that you wanted me to point some out. I googled to get the first and used the search tool above to get the others.
Author
Don't want to be too confrontational, but based on my limited knowledge, I put in the same search terms as you did and got similar results.
None of which were any use as all these variants on negamax state in paragraph #1 that they're for two player games, whilst I've stated several times that my game has seven players...
None of which were any use as all these variants on negamax state in paragraph #1 that they're for two player games, whilst I've stated several times that my game has seven players...
Quote:Original post by Rubicon
...whilst I've stated several times that my game has seven players...
Which was not in your original post. You wanted links to negamax and brute force. If you had known ahead of time that negamax and brute force were not going to work for your 7-player game, why ask about them?
Any algorithm that is typically good for chess/checkers/go, etc. is going to fall apart for that many people because the search space gets ridiculous.
The thing I tell all people is to look at your game from the point of view of the player. What things do you take into account when you make your move? You are not likely to look too many moves ahead in a situation such as that. You are simply doing something that seems like it maximizes your immediate result while minimizing the potential to get pwnd. While that seems, at its core, like a tree-based search (e.g. minimax et al), you can also do it by simply scoring certain factors. All of this is game-dependent, of course, so further speculation about how to approach it is a problem. But that's what you are going to find in books anyway. You will have to adapt.
Thanks for the books link, i'm also interested to read them.
And InnocuousFox, i agree with your points:look at your game from the point of view of the player.
And InnocuousFox, i agree with your points:look at your game from the point of view of the player.
This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement