8.1 What is AI Planning?
Planning involves finding a sequence of actions to achieve a goal from an initial state.
Key Characteristics:
- Requires representation of states, actions, and goals
- Actions have preconditions and effects
- Deals with temporal reasoning and constraints
8.2 Planning Domain Definition Language (PDDL)
Standard language for representing planning problems with:
- Domain file: Defines predicates and actions
- Problem file: Defines objects, initial state, and goal
(define (domain blocksworld)
(:requirements :strips)
(:predicates (on ?x ?y) (clear ?x))
(:action move
:parameters (?from ?to)
:precondition (and (on ?from ?to) (clear ?from))
:effect (and (not (on ?from ?to)) (on ?from ?to))))
8.3 Planning Algorithms
| Algorithm | Description |
|---|---|
| STRIPS | Classical planning with add/delete lists |
| Partial-Order Planning | Flexible ordering of plan steps |
| GraphPlan | Creates planning graph to find solutions |
| HTN Planning | Hierarchical decomposition of tasks |
8.4 Frequently Asked Exam Questions
- Compare classical planning with problem solving.
- Write PDDL for a simple planning domain.
- Explain how STRIPS represents actions.
- What are the advantages of HTN planning?
- Solve a blocksworld problem using planning techniques.