I am trying to create a 2-player Atlas game bot along the lines of chess bots by using the Minimax algorithm. I am using a database of just countries and capitals to induce strategy in the game. There are 2 ways the game could end :
- The player doesn't know the place name with that starting letter
- There are no places left with that starting letter
The first way could be implemented by placing a timer. The second way is quite interesting, we can create a decision tree where each node represents the available moves for each player.
But a significant problem is unlike chess, where the player and opponent both have their resources, and the evaluation function is calculated using those resources, after which the algorithm tries to maximize the player's route and minimize the opponent's. Here, both players have the same set from which they have to choose, and I don't think minimax can work well here. Will negamax work or is genetic algorithms the only option or maybe I do not need these fancy algorithms?
