Our aim is to train an artificial intelligence player for Abalone game based on Alpha Zero's algorithm.
Abalone is an award-winning two-player abstract strategy board game. Players are represented by opposing black and white marbles on a hexagonal board with the objective of pushing six of the opponent's marbles off the edge of the board. Hexagonal board, more possible actions and more complex rules make this game harder to implement than Go.
We want to:
test the generality of Alpha Zero's algorithm on more complex issues,
see whether our AI player can discover some remarkable Abalone game knowledge during
its self-play training process,
compare the performance of Alpha Zero's algorithm with other algorithms.
In order to prevent the game from being so complicated that the training process cannot be completed within three months, we have looked for other games as alternatives, such as Oware.
Related work:
Silver, D., Schrittwieser, J., Simonyan, K., Antonoglou, I., Huang, A., Guez, A., ... & Chen, Y. (2017). Mastering the game of go without human knowledge. Nature, 550(7676), 354.
Ozcan, E., & Hulagu, B. (2004). A simple intelligent agent for playing abalone game: Abla. In Proc. of the 13th Turkish Symposium on Artificial Intelligence and Neural Networks (pp. 281-290).
Lemmens, N. P. P. M. (2005, June). Constructing an abalone game-playing agent. In Bachelor Conference Knowledge Engineering, Universiteit Maastricht.
Our aim is to train an artificial intelligence player for Abalone game based on Alpha Zero's algorithm.
Abalone is an award-winning two-player abstract strategy board game. Players are represented by opposing black and white marbles on a hexagonal board with the objective of pushing six of the opponent's marbles off the edge of the board. Hexagonal board, more possible actions and more complex rules make this game harder to implement than Go.
We want to:
In order to prevent the game from being so complicated that the training process cannot be completed within three months, we have looked for other games as alternatives, such as Oware.
Related work: Silver, D., Schrittwieser, J., Simonyan, K., Antonoglou, I., Huang, A., Guez, A., ... & Chen, Y. (2017). Mastering the game of go without human knowledge. Nature, 550(7676), 354.
Ozcan, E., & Hulagu, B. (2004). A simple intelligent agent for playing abalone game: Abla. In Proc. of the 13th Turkish Symposium on Artificial Intelligence and Neural Networks (pp. 281-290).
Lemmens, N. P. P. M. (2005, June). Constructing an abalone game-playing agent. In Bachelor Conference Knowledge Engineering, Universiteit Maastricht.