Goals and State Space

• Goals are good old fashioned AI. Newell and Simon explored goals and subgoals with the General Problem Solver in 1960.
• That means a lot of the problems have been sorted or at least considered.
• Goaling fits into State Space search really well.
• There are two ways to search. The agent can plan, or it can actually make actions.
• That is, the agent can explore the state space via simulation or via action (moving about the environment).
• Goals and subgoals can help here to give you intermediate states.