Extending the minimax algorithm for multiple opponents

Question

The minimax algorithm is well described for two players for games like tic-tac-toe. I need to write an AI for a Tank game. In this game, the tanks have to move in a maze that has obstacles in the form of walls. The goal is to collect coin piles. If it was only two players the minimax algorithm can be implemented. But how to implement it for more than two? As at each turn, each player will try to maximize his own winning edge. I can not think of all the players as one enemy trying to reduce only my winning edge creating the two-player levels as in the original minimax algorithm. Please excuse me if the question is not in a good format. Still new to this forum.

Answer 1

What you are trying to implement is very hard. You cannot use the minimax for this implementation. You can create algorithms that can able to learn on a strategical level what needs to be done. Transform the game into a two-player one: me vs. the others and start from here.