State Space
- I think
State Space is a crucial concept for AI.
- The problem you're trying to solve is broken up into a
series of discrete states.
- Not all problems can be described this way, but many can, and
all can be approximated this way.
- You also describe the transitions between states. How do
you go from one to another.
- You frequently have start and goal states.
- The problem then becomes a question of search.
- How can you move from the start to the goal.
- For example, how can you describe a maze as a state space?
- How can you search to find a path from the start to the goal.
- From the graph lecture, you can
use Depth First Search or Breadth First Search.
- Also note that the size and shape of the state space tells you
a lot about what to do.
- If it is huge, you can't use exhaustive techniques, unless it's
reqular.
- How would you describe a tic-tac-toe game in state space?