6.7 Review

The following are the main points you should have learned from this chapter:

  • Planning is the process of choosing a sequence of actions to achieve a goal.

  • An action is a partial function from a state to a state. Two representations for actions that exploit structure in states are the STRIPS representation, which is an action-centric representation, and the feature-based representation of actions, which is a feature-centric representation.

  • Planning algorithms can be used to convert a planning problem into a search problem.

  • A forward planner searches in the state space from the initial state to a goal state.

  • A regression planner searches backwards from the goal, where each node in the search space is a subgoal to be achieved.

  • A planning problem for a fixed horizon can be represented as a CSP, and any of the CSP algorithms can be used to solve it. The planner may need to search over horizons to find a plan.

  • A partial-order planner does not enforce an ordering between actions unless there is a reason to make such an ordering.

  • The video game industry has successfully exploited AI planning techniques.