SHOP 2 is a hierarchical task network (HTN) planner developed in LISP that was awarded for its performance in the 2002 international planning competition. It improves upon SHOP by allowing partially ordered tasks and subtasks, and supports many PDDL features. SHOP 2 decomposes complex tasks into subtasks through methods until reaching primitive tasks. It has been applied to domains like video games, evacuation planning, evaluating terrorist threats, and automated composition of web services.

1. SHOP 2
Simple Hierarchical Ordered
Planner 2
slides by Andrea Tucci

2. ● Introduction
● Hierarchical Task Networks
● Features
● SHOP 2 Algorithm
● Using SHOP 2
● Applications (SHOP & SHOP 2)
● References
Overview

3. ● Introduction
● Hierarchical Task Networks
● Features
● SHOP 2 Algorithm
● Using SHOP 2
● Applications
● References
Overview

4. Introduction
Simple Hierarchical Ordered Planner 2 (SHOP 2) :
● Based on Hierarchical Task Network (HTN)
● Implemented in LISP
● Developed by D.Nau, T.Au, O.Ilghami, U.Kuter, D.Wu,
and F.Yaman, H. Muñoz-Avila, J.W. Murdock in 2001;
● Award-winning at International Planning Competition
(2002) for its performance

5. Introduction (2)
● Steps of a plan have the same order as their execution
(like SHOP).
>> this speed up reasoning process
● SHOP 2 characteristics go further :
○ task and subtask can be partially ordered;
○ many PDDL features included (e.g. quantifiers,
conditional effects)
○ possibilities to satisfy a method's precondition are
sorted according to specifications provided by the
programmer
○ others

6. ● Introduction
● Hierarchical Task Networks
● Features
● SHOP 2 Algorithm
● Using SHOP 2
● Applications
● References
Overview

7. Hierarchical Task Networks
(HTNs)
● Produces actions to reach a (complex) task.
● Task can be decomposed in subtasks, primitive or
complex
● Method : how a complex task can be decomposed
● Planner decomposes complex tasks until a primitive one
is found

8. Hierarchical Task Networks
(HTNs) - (2)

9. Hierarchical Task Networks
(HTNs) - (3)
Planner :
● Choose a method for decompose a complex task into
subtasks
● Repeat until a plan is found or we came out with an
infeasible plan
● In the latter case, backtrack and try other implementing
methods.

10. Hierarchical Task Networks
(HTNs) - (4)
Example :
(transport-two p1 p2)
initial state :
(package p1), (at p1 l1), (destination p1
l3), (available-truck t1), (at t1 home)
(package p2), (at p2 l2), (destination p2
l4), (available-truck t2), (at t2 home)

11. Hierarchical task Networks
(HTNs) - (4)

12. ● Introduction
● Hierarchical Task Networks
● Features
● SHOP 2 Algorithm
● Using SHOP 2
● Applications
● References
Overview

13. Features - Basic Elements
Specify a domain description that consists of a set of
methods, operators and axioms.
>Syntactical Note
A variable is preceded by "?" symbol.
A comment is preceded by ";".
● Tasks
Action to be done. They can be either primitive or
complex.
(s t1 t2 ... tn)
A primitive task is preceded by "!" symbol.
● Operators
Indicates how a primitive task can be executed.
(:operator h P D A [c])

14. Features - Basic Elements - (2)
● Methods
How to decompose a complex task.
(:method h [n1] C1 T1 [n2] C2 T2 ... [nk] Ck
Tk)
● Axioms
Used to infer precondition that are not explicitly
asserted in our current state (e.g. including quantifiers,
implications etc).
(:- a [name1] E1 ... [namen] En)
(:- head tail)
Which means that head is true if tail is true (Horn
clauses).

15. Features - Basic Elements 6
● External Function Call
Result is returned by a procedure, e.g
numeric evaluation.
( call f t1 t2 ... tn )
where f is a function symbol

16. Features - Additional
● Sorting variable bindings
● Branch and Bound Optimization
● PDDL Operator Translation
● Protected Condition and Anti-Interleaving
● Debugging Facilities

17. ● Introduction
● Hierarchical Task Networks
● Features
● SHOP 2 Algorithm
● Using SHOP 2
● Applications
● References
Overview

18. SHOP 2 Algorithm
if t is primitive, use an instance a of an
operator that matches t and has satisfied
precondition in s. Apply a.
if t is complex, find a method that can be
applied to t and decompose it.

19. ● Introduction
● Hierarchical Task Networks
● Features
● SHOP 2 Algorithm
● Using SHOP 2
● Applications
● References
Overview

20. Using SHOP 2
● SHOP 2 requires lisp
● JSHOP also available, based on Java
Create a program in SHOP 2
1. Define a domain
2. Define one (or more) problems
3. run!

21. Using SHOP 2 - (2)
; extremely simple example
(defdomain basic (
(:operator (!pickup ?a) () () ((have ?a)))
(:operator (!drop ?a) ((have ?a)) ((have ?a)) ())
(:method (swap ?x ?y)
((have ?x) (in-range ?x ?y))
((!drop ?x) (!pickup ?y))
((have ?y) (in-range ?x ?y))
((!drop ?y) (!pickup ?x))))
)
; axiom
(:- (in-range ?x)
yes
((position ?x ?p1) (position ?y ?p2) (max-range ?r) (call >= ?r
(call - ?p2 ?p1)))
)
)
(defproblem problem basic
((have phone) (max-range 20) (position phone 5) (position key 10))
((swap phone key)))
PLAN -> (!drop phone) (!pickup key)

22. ● Introduction
● Hierarchical Task Networks
● Features
● SHOP 2 Algorithm
● Using SHOP 2
● Applications
Overview
○ Video games
○ Evacuation
○ Evaluation of terrorist threats
○ Automated composition of web services
○ ...more.
● Resources

23. Application - Video games

24. Application - Video Games (2)
● The use of HTNs in video games arise in
the last years.

25. Application - Video Game (3)
● An example : Killzone 2 (Guerrilla Games, PS3 2009)
● First Person Shooter (FPS)
● Planner based on SHOP syntax
● Killzone AI structure : goal driven
● Behavior associated with the selected goal

26. Application - Video Games (4)
● Killzone 2 AI : HTN-based
● High-level rules -> what to do
● Low-level rules -> how to do selected task(s)
● More flexibility
● HTN also control squads

27. Application - Video Games (5)
example
(:method (select_weapon_and_attack_as_turret ?inp_threat)
(branch_use_bullets // Only use bullets against humanoids and turrets.
(and
(or (threat ?inp_threat humanoid) (threat ?inp_threat turret) )
(distance_to_threat ?inp_threat ?threat_distance)
(call lt ?threat_distance @weapon_bullet_max_range) )
((attack_as_turret_using_weapon_pref ?inp_threat wp_bullets))
)
( branch_use_rockets // Don't use rockets against humanoids and turrets.
(and
(not (threat ?inp_threat humanoid)) (not (threat ?inp_threat
turret))
(distance_to_threat ?inp_threat ?threat_distance)
(call lt ?threat_distance @weapon_rocket_max_range) )
((attack_as_turret_using_weapon_pref ?inp_threat wp_rockets))
)
)

28. Application - Video Games (6)
some numbers
Individual bot domain
● 360 methods
● 1048 branches
● 138 behaviors
● 147 continue branches
During multiplayer game (14 bots / max. 10 turrets / max. 6
drones / squads)
● Approx. 500 plans generated per second
● Approx. 8000 decompositions per second
● Avg. 15 decompositions per planning.
● Approx 24000 branch evaluations per second.

29. Application - Evacuation
● US Naval Research Laboratory’s Hierarchical
Interactive Case-Based Architecture for Planning
(HICAP) helps human planners to build evacuation
plans.
● Generative and case-specific planning. SHOP provides
the generative component.
● Interactive plan modification : human expert can modify
tasks and plans interactively.

30. Application - evaluating terrorist
threats
● Naval Research Laboratory’s Analogical
Hypothesis Elaboration for Activity
Detection (AHEAD) helps to understand
and evaluate terrorists threats.
● A HTN domain encode hostile activities.
● AHEAD takes a hypothesis as input and
returns a model of the hostile activity.

31. Application - evaluating terrorist
threats
● SHOP 2 produces a plan compatible with
hypothesis.
● SHOP 2 queries an external evidence
database to make some tests

32. Application - automated
composition of web services
● Web service : a part of web that offer
functionality
● The OWL-S (Web Ontology Language for
Services) is a language for semantic markup
of Web services
● It describes services as complex or atomic
processes with preconditions and effects.
● University of Maryland has developed an
algorithm to translate OWL-S process-
models to SHOP 2 methods and operators.

33. Application - automated
composition of web services (2)
● Goal : compose web services
● User can compose them by using SHOP
2
● Result : HTN plan that corresponds to the
composite process.

34. Application - others
● Fighting forest fires
● Material selection for manifacturing
● Project Planning
etc
Check this paper out SHOP & SHOP 2
applications

35. ● Introduction
● Hierarchical Task Networks
● Features
● SHOP 2 Algorithm
● Using SHOP 2
● Applications
○ Video games
○ Evacuation
○ Location-based services
○ Automated composition of web services
○ ...more.
● References
Overview

36. References
● SHOP 2 : An HTN Planning System - (D.
Nau et al, 2003)
● A hierarchically-layered multiplayer bot
system for a first-person shooter (Tim
Verweij, 2007)
● Killzone 2 Multiplayer Bots - ( Alex
Champandard, Tim Verweij, Remco
Straatman, 2009)
● Application of SHOP and SHOP 2 - (D.Nau
et al, 2005)
● Documentation for SHOP 2

37. Thanks for your attention
Andrea Tucci
@andreatux