[PDF]

Multi-player Othello game with AI and networking


Kieran Chocian

20/10/2022

Supervised by Frank C Langbein; Moderated by Shancang Li

Othello is a turn-based game on an 8x8 grid that consists of white and black discs. Each player places a disc that outflanks one or more of their opponent discs, where every disc outflanked is flipped. The game ends when neither player can place a disc, where the aim is to have more discs left on the board than your opponent does.

This project aims to create a networked server that allows a user to play Othello against either another human in real time, or against multiple AI agents of varying difficulty. These agents should be able to play on multiple board sizes, and also aim to answer questions such as "How does the number of concurrent users affect the performance of an AI agent?" and "What attributes of a convolutional neural network lead to the best prediction of an Othello move?".


Final Report (20/10/2022) [Zip Archive]

Publication Form