This project creates an AI agent to play games of Tic-tac-toe.
The included my_agent
function should be modified to return an appropriate move for a given board position. Additional implementation details are described in play.py
.
Tic-tac-toe is a short, perfect information game with a known optimal solution. Perfect play results in every game ending in a draw.
On modern hardware, exhaustive search using minimax should be possible to achieve optimal play. Other approaches, such as Monte Carlo Tree Search should also produce good results while using much less compute.