MiniMax algorithm with alpha-beta prunig

[Guest_imported]

Hello,

my question concerns not the gist of the algorithm. In order to play against a computer player the computer must have an approach to calculate the next move.

I must programme the TWIXT-game. This is a board game with 24*24 matrixes. In each of them you can stick a peg. Two players can play this game only. One player tries to get from west to east and the other one from north to south.

The pegs must be stuck in a knight's move. In this case the pegs are connected with a so-called wall. The winner is that player who has finished a continuous line from one side to the other side.

Ok, this was just to get the point of my question. The minimax algorithm needs a situation analyzer that assesses each possible move.

My question: Do you have any idea which criteria would be the best for assessing each possible move?

The better the criteria is, the stronger the computer playz.

Any idea would be highly appreciated.

Great thanks in advance.

woody_willow

 

[cs1]

Can't the distance of the wall developed thus far be a deciding factor? You can also count the number of moves made for that...
Have you read Elain Rich & Kevin Knight's "Aritificial Intelligence" book? It has an elaborate explanation on alpha-beta cutoffs.

- Karthika

 

[Guest_imported]

Hello,

couldn't get what you mean. Can you be more concrete please.

I haven't read Elain Rich & Kevin Knight's "Aritificial Intelligence" already.

Do you have some experience in this point?

woody_willow

 

[cs1]

I have no previous knowledge of this game. The Artificial Intelligence book that I mentioned has detailed explanation of strategies used in several games, how to compute the alpha-beta cutoffs... You will get many ideas by reading it.
The 2 factors that I mean are:
1. The displacement of the player from the start position.
2. The number of moves made till now.
- Karthika