1. 1 USC CS 541 AI Planning Lecture Notes Yolanda Gil Plan Representation and Reasoning with Description Logics and Ontologies Yolanda Gil Lecture Notes, October 4, 2000 CS 541 Artificial Intelligence Planning www.isi.edu/~gil/cs541

2. 2 USC CS 541 AI Planning Lecture Notes Yolanda Gil Outline Representing actions and plans with description logic  Action taxonomies (CLASP)  Plan taxonomies (SUDO-PLANNER)  Goal taxonomies (EXPECT) Planning ontologies  Process Specification Language (PSL), NIST  PLANET

3. 3 USC CS 541 AI Planning Lecture Notes Yolanda Gil Representing Knowledge in Description Logic (DL) Description logics are extensions of frame-based systems where classes can be defined intensionally  Ex: SUVs are vehicles with 4 seats that weight between 1T and 2T Class taxonomy is automatically generated through subsumption  A subsumes B iff all instances of B are also instances of A Instances can be automatically classified  Ex: MyNewCar is a vehicle with 4 seats that weighs 1.3T => MyNewCar is an SUV Relations can also have definitions and can be classified Tradeoff between expressivity and efficient reasoning Some well know description logic systems: CLASSIC, LOOM, NIKL

4. 4 USC CS 541 AI Planning Lecture Notes Yolanda Gil Representing Planning Knowledge in Description Logics: Overview Action taxonomies in CLASP  extended language to represent action networks Plan taxonomies in SUDO-PLANNER  plan subsumption of partially ordered plans Goal taxonomies in EXPECT  expressive representations of goals and their parameters These systems can exploit the descriptions of all the objects in the domain ( domain knowledge ) in order to reason about action, goal, and plan descriptions

5. 5 USC CS 541 AI Planning Lecture Notes Yolanda Gil CLASP: CLAssification of Scenarios and Plans [Devanbu and Litman 94] Extension of a DL system (CLASSIC)  Language to express action networks –Sequence, loop, repeat, test, subplan  Subsumption and classification algorithms for that language –Action network subsumption viewed as DFA acceptance Propositional, STRIPS-style representation of actions  States (goals are represented as states)  Actions  Plans  Scenarios (plan instances) Reasoning based on these descriptions:  Organizing plan classes  Retrieving plan types  Validation of scenarios

6. 6 USC CS 541 AI Planning Lecture Notes Yolanda Gil Core Classes in CLASP (DEFINE-CONCEPT Action (PRIMITIVE (AND Classic-Thing (AT-LEAST 1 Actor) (ALL ACTOR Agent) (EXACTLY 1 PRECONDITION) (ALL PRECONDITION State) (EXACTLY 1 ADD-LIST) (ALL ADD-LIST State) (EXACTLY 1 DELETE-LIST) (ALL DELETE-LIST State) (EXACTLY 1 GOAL) (ALL GOAL STATE)))) (DEFINE-CONCEPT State (PRIMITIVE Classic-Thing)) (DEFINE-PLAN Plan (PRIMITIVE (AND Clasp-Thing (EXACTLY 1 INITIAL) (ALL INITIAL State) (EXACTLY 1 GOAL) (ALL GOAL State) (EXACTLY 1 PLAN-EXPRESSION) (ALL PLAN-EXPRESSION (LOOP Action)))))

7. 7 USC CS 541 AI Planning Lecture Notes Yolanda Gil Defining Actions, States and Plans in CLASP in a Telephony Domain (DEFINE-CONCEPT System-Act (AND Action (ALL ACTOR System-Agent))) (DEFINE-CONCEPT Connect-Dialtone-Act (AND System-Act (ALL PRECONDITION (AND Off-Hook-State Idle-State)) (All Add-LIST Dialtone-State) (ALL DELETE-LIST Idle-State (ALL GOAL (AND Off-Hook-State Dialtone-State)))) (DEFINE-CONCEPT Callee-Off-Hook-State (PRIMITIVE State)) (DEFINE-CONCEPT Callee-On-Hook-State (PRIMITIVE State)) (DEFINE-CONCEPT Callee-Off-Caller-On-State (AND Callee-Off-Hook-State Caller-On-Hook-State)) (DEFINE-PLAN Pots-Plan (AND Plan (ALL PLAN-EXPRESSION (SEQUENCE (SUBPLAN Originate-And-Dial-Plan) (TEST (Callee-On-Hook-State (SUBPLAN Terminate-Plan)) (Callee-Off-Hook-State (SEQUENCE Non-Terminate-Act Caller-On-Hook-Act Disconnect Act))))))) (DEFINE-PLAN Originate-And-Dial-Plan (AND Plan (ALL PLAN-EXPRESSION (SEQUENCE Caller-Off-Hook-Act Connect-Dialtone-Act Dial-Digits-Act))))

8. 8 USC CS 541 AI Planning Lecture Notes Yolanda Gil Defining Instances in CLASP (CREATE SCENARIO pots-busy-scenario (AND Plan (FILLS INITIAL state-u1on-u2off) (FILLS GOAL state-u1on) (FILLS PLAN-EXPRESSION (caller-off-hook-u1 connect-dialtone-on-u1 dial-digits-u1-to-u2 non-terminate-on-u2 caller-on-hook-u1 disconnect-u1)))) (CREATE-IND state-u1on-u2off (AND state-U1on State-U2off)) (CREATE-IND connect-dialtone-on-u1 (AND Connect-Dialtone-Act (FILLS ACTOR switching-system) (FILLS PRECONDITION state-u1off-idle)))

9. 9 USC CS 541 AI Planning Lecture Notes Yolanda Gil SUDO-PLANNER [Wellman 88] Exploits subsumption to control the search during plan generation Actions represented in DL (NIKL), organized in taxonomy Plans represented as partially ordered sets of actions  Eliminate search nodes whose plan is subsumed ( dominated ) by other nodes SUDO-PLANNER had other features not discussed here:  Uncertainty reasoning and partial goal satisfaction  Policy constraints that relate actions to external events  Conditional effects  Qualitative probabilistic networks

10. 10 USC CS 541 AI Planning Lecture Notes Yolanda Gil Action Taxonomy in SUDO-PLANNER (defconcept surgery :is (:and action (:the route invasive-path-into-body))) (defconcept biopsy :is-primitive action...)) (defconcept open-lung-biopsy :is (:and biopsy (:the route open-lung-path))) (defconcept open-lung-path :is (:and invasive-path-into-body...)) System deduces that open-lung-biopsy is a surgery

11. 11 USC CS 541 AI Planning Lecture Notes Yolanda Gil Plan Representation and Subsumption in SUDO-PLANNER Plan is described as a set of action types associated with identifiers  [(surgery, id1) (CABG, id2)] Plan is simplified if action subsumption and same id  [(surgery, id1) (CABG, id1)] -> [(surgery, id1)] Plan subsumption  Action network viewed as bipartite graph matching a1 a2 a5 a3 a4 a6 a1 a4 a5 a2 a3 a6

12. 12 USC CS 541 AI Planning Lecture Notes Yolanda Gil A* a1 A* a2 A* a2 b7 A* A* a1 b5 A* a1 b5 A*... X A* = {ai…aj} ai subsumes aj when i<j Eliminating Redundant Paths in Plan Space Search Dominance-based planning :  Generate new nodes by adding constraints to search nodes  Derive dominance (i.e., subsumption) based on domain knowledge  Eliminate nodes in the plan graph that are dominated by others

13. 13 USC CS 541 AI Planning Lecture Notes Yolanda Gil Reasoning about Goals in EXPECT [Swartout et al 98] Highly declarative representation of goals  Goals as verb-based expressions  Rich language of goal parameter types – Qualification parameters that describe the type of task –Intentional and extensional sets Given a goal, matcher looks for methods (operators) that have the capability of achieving that goal  can match variabilized goals  can decompose goal into subgoals through reformulations Goal representations have been used in several contexts:  representing planning goals  problem solving  agent matchmaking

14. 14 USC CS 541 AI Planning Lecture Notes Yolanda Gil Representing Goals in EXPECT Represented as a case grammar (verb + roles)  ex: ESTIMATE OBJ duration OF trip Roles can be filled by:  a specific instance: add OBJ 3 TO 5  a concept: compute OBJ (spec-of factorial) OF 7  a type of instance: divide OBJ number BY 2  extensional sets: find OBJ (spec-of maximum) OF (54 15 256)  intensional sets: add OBJ (set-of number) find OBJ (set-of (spec-of violated-constraint)) IN configuration Roles filled by concepts express task qualification parameters declaratively (compute-factorial ?n) -> (compute (obj (spec-of factorial)) (of number)))

15. 15 USC CS 541 AI Planning Lecture Notes Yolanda Gil Matching Goals in EXPECT Desired goals and available capabilities are automatically translated to LOOM concepts Classifier is used to find most specific method capability that subsumes the posted goal Self-organizing method taxonomy move cargo aircraft OBJ WITH move cargo truck OBJ WITH move cargo vehicle OBJ WITH move cargo C-140 OBJ WITH Goal: (move (OBJ (inst-of cargo)) (WITH C-140)) Method capability: (move (OBJ (inst-of cargo)) (WITH (inst-of aircraft)))

16. 16 USC CS 541 AI Planning Lecture Notes Yolanda Gil Flexible Matching through Goal Reformulation When no capability matches a posted goal, but more specific versions of the goal match  ex: no method to estimate round-trip time (rtt) of a vehicle, but there are methods to estimate rtt of aircraft and trucks Use descriptive knowledge to reformulate goal  reexpress goal into subgoals by breaking down one of the arguments  recombine the results of solving subgoals Conjunctive (disjunctive) subgoals produce conjunctive (disjunctive) reformulations Types of reformulations  Covering reformulation: subgoals cover partitions of a class  Set reformulation: subgoals iterate over elements of a set  Input reformulation: subgoals handle each of the subtypes

17. 17 USC CS 541 AI Planning Lecture Notes Yolanda Gil Find route from location1 to location2 Find egress route from Ryad to Kuwait city Calculate RTT for transport aircraft Calculate round-trip time (RTT) for aircraft Calculate RTT for combat aircraft A) Subsumption-based match: the posted goal is subsumed by a capability B) Reformulation-based match: the posted goal can be satisfied by combining two or more existing capabilities Find route from city1 to city2 Find route from location1 to location2 C) Reverse subsumption-based match: a capability can satisfy some aspect of the goal Find addresses of US citizens in Kuwait Find phone numbers of US citizens in Kuwait D) Partial match: a capability is similar/related to the posted goal Goal Matching in EXPECT

18. 18 USC CS 541 AI Planning Lecture Notes Yolanda Gil Overview of Planning Ontologies Why planning ontologies  knowledge reuse  knowledge sharing  knowledge modeling Process descriptions in PSL  temporal constraints  resources Describing plans in PLANET  can represent state-based, plan-based search, hierarchical plans  captures plan representations understandable by people

19. 19 USC CS 541 AI Planning Lecture Notes Yolanda Gil Process Specification Language (PSL) [NIST, 99] National Institute of Standards and Technology (NIST), Manufacturing Systems Division  Academic and industrial collaborators Proposed to Int’l Standards Organization (ISO) PSL core represents widely accepted commitments  activity, activity-occurrence, object, timepoint PSL extensions accommodate possible shareable agreements Contains axioms defining terms and constraints Available at http://www.mel.nist.gov/psl/

20. 20 USC CS 541 AI Planning Lecture Notes Yolanda Gil PSL Overview

21. 21 USC CS 541 AI Planning Lecture Notes Yolanda Gil PSL Modules for Activities and Orderings

22. 22 USC CS 541 AI Planning Lecture Notes Yolanda Gil PSL Core Activity  Generic activity: occurrences, interruptions, nondeterministic, subactivities  Ordering: ordering over activities, complex ordering relations, junctions Objects  Resources: capacities, homogeneous sets, inventories, divisibility, usage, resource paths, pools, requirements, resource roles, substitutability  States: defined, constraints Timepoints  Duration theory, activity durations, temporal orderings

23. 23 USC CS 541 AI Planning Lecture Notes Yolanda Gil PLANET: a PLAN Semantic nET [Gil & Blythe 00] Capture unifying views on planning algorithms  constraints, commitments, task templates, alternative choices  state-based and objective-based goals  operator-based and HTN-based plans Represent manually created plans  typically include unintended flaws (incomplete, unjustified, inconsistent) Capture planning context  initial constraints (guide, user advice, preferences) and restraints  initial state, constraints and goals may be incompatible Available from http://www.isi.edu/expect/projects/planet/

24. 24 USC CS 541 AI Planning Lecture Notes Yolanda Gil Some Terms Defined in PLANET Planning problems  Planning problem context: world state, desired goals, external constraints  Planning problem: candidate plans (rejected, feasible, selected) Goals and effects  Goals: state-based goals, objective-based goals  Human readable descriptions Actions, operators, and tasks  Plan task descriptions: plan task templates, plan tasks,  Capabilities, preconditions, effects, subtasks, primitive tasks, plan steps Plans  Commitments, sub-plans, planning level

25. 25 USC CS 541 AI Planning Lecture Notes Yolanda Gil PLANET: An Ontology for Representing Plans justified complete consistent feasible plan-commitments plan-refinements sub-plans capability effects preconditions planning-level human-readable description sub-tasks task-of Plan-task-template task-template commitments accomplishes ordering temporal planning-problems initial-state world-state desired-goals external constraints candidate-plans rejected feasible selected unexplored planning - level state-based -goal-spec objective-based -goal-spec resource-needed amount when-needed Plan-task-description Plan-task Plan Planning-problem-context Goal-specification Resource- requirements