SnookAI

Abstract

Billiards automation has been a dedicated research field in computer science that has motivated computer scientists to create systems that address this problem. This dissertation discusses the artificial intelligence involved in a game of snooker. An A.I. is to be implemented using researched algorithms to select the very best possible shot for every state of the game that the computer is faced with (whether it is an attacking shot or a defensive shot). Adding to this, it should be capable of handling various game situations within a reasonable amount of time. Different attacking strategic methods will be researched and implemented in order to see which performs the best in clearing all the balls of a snooker table. On the other hand, algorithms which deal with the defensive standpoint of the game are also implemented so that the current player minimises the alternatives of the opponent for attacking moves. Finally, we will have an A.I. which makes use of these researched algorithms to deal with whichever table state the computer player is faced with. In order to win a game of snooker, the player who is faced with a particular game situation is required to search for the possible attacking shots. After choosing an attacking shot, the player must ensure that his break is long lived by landing the cue ball in an ideal position which leads to other attacking shots. This goes on until the table is cleared from all the balls. The A.I. will pretty much be making the same decisions during his turns. The latter should keep the consistency of potting balls and control the cue ball for positional play. This consistency is backed with suitable algorithms as a product of the contribution of our research.