Today we started to design our algorithm. We began by discussing the Wumpus World problem and how to approach it. We knew that there wouldn't be just one "good" way to solve the problem, so we approached it from many different angles.
We reviewed the simulation program and learned how it was designed. While doing so, we were able to narrow down the number of ways our logic could be executed in order to optimize our algorithm. We reasoned that it was optimal to use everything that is provided by the program itself in order for the Agent to be efficient. We decided that we would keep the program's automatically generated output as a continuously-updated list of senses so that the Agent could analyze the senses it receives in real-time.
On the first day of the project, we also made sure that we had at least one environment that worked between the three of us - Anthony confirmed that his simulator worked perfectly.
We still have some things to iron out, so we probably won't start coding on Day 2.
Our notes:
Comentarios