| Abstract | This research analyzes artificial intelligence techniques for Konane. The game Konane, also known as Hawaiian checkers, is a two-player, zero-sum strategy board game ideally suited for this type of research. In order to have a successful strategy, a player must consider many future possibilities. We compare computing agents that use informed and uninformed searching algorithms but focus our investigation on the effectiveness of the minimax algorithm. By altering variables such as the cutoff depth for searching the game tree and incorporating alpha-beta pruning, we begin to see varying levels of success and efficiency from the competing computing agents. The outcome of this research is an analysis of the effectiveness of each computing agent showing the positive correlation between the depth of the search tree and the percentage of games won and the exponential relationship that exists between the number of nodes explored and the depth of the search tree. |
