For our decision theory project, we decided to build an artificial intelligence system in the form of a rational agent for a minesweeper game. We used a value iteration approach that created an abstract policy with which our agent wins on average 3 to 4 times more games than a random agent.