The game “Knightthrough” is a simple board game, played on a regular chess board, only with pieces that can move in the same way as a Knight can in chess. The player who manages to reach the opposing side of the board first wins the game. In this project (part of a course in the first year Master AI program at Maastricht University), the goal was to implement an agent to play this game, using the Alpha-Beta search algorithm as the core of the agent.
The agent can be considered a fairly strong agent. On top of the regular Alpha-Beta search algorithm, it uses enhancements such as Aspiration Search, a Transposition Table, and the Killer Move Heuristic. It also uses an efficient bitboards-based state representation. Of course, there are still many more possible improvements in the literature, so this is not completely state-of-the-art.