AI Mancala

JOHNATHAN A. SAUNDERS, and JATIN BHAKTA

The purpose of our project is to create an artificially intelligent program that will be capable of playing the game Mancala. Mancala is a game with a board that has 6 “holes” on each row and one larger hole on each end. Mancala starts off with 4 “stones” in each “hole” and the first player will choose any pile on his/her side and put one stone in the consecutive piles in a counter-clockwise manner. The player’s turn ends when he/she lands in an empty hole. After no stones are left to choose from, the player with the most stones in their end side wins the game. The artificial intelligence is implemented using the minimax algorithm. A user inputs a move to change the current board state. Then the AI assesses the board state every turn and returns a semi-optimal move choice for itself. The purpose of this project is to compare the computers performance when playing Mancala. There are plans to compare the AI’s performance against other AIs such as some machine learning algorithms - for example a neural network.