In the class CS-4341, Artificial Intelligence, we were tasked to build an agent that could play the game of Gomoku. In order to achieve this the team implemented the MiniMax algorithm, Alpha-beta-pruning as well as our own understanding of evualtion function to facilitate the previous two algorithms. Additionally to aide in the agents compentency the team built, from scratch, a linear neural network.
Where the number indicates the threat size and dead indicates the move is not preventable (i.e. there are two or more places a stoned can be placed such that it will complete the move). A total of 20 shapes are possible with the seven above categories.
Potential eval function
7 categories of stone positions:
Where the number indicates the threat size and dead indicates the move is not preventable (i.e. there are two or more places a stoned can be placed such that it will complete the move). A total of 20 shapes are possible with the seven above categories.