Practical Hierarchical Task Network (HTN) Planning Examples of using - - PDF document

▶
practical hierarchical task network htn planning examples
SMART_READER_LITE
LIVE PREVIEW

Practical Hierarchical Task Network (HTN) Planning Examples of using - - PDF document

Nov 21, 2022 433 likes •732 views

Practical HTN Planning Putting HTN Planning into Use Practical Hierarchical Task Network (HTN) Planning Examples of using the core HTN technique along with other essential planning, search and constraint satisfaction technologies to realise

slide-1
SLIDE 1

1

Practical HTN Planning

Putting HTN Planning into Use

Practical Hierarchical Task Network (HTN) Planning Examples of using the core HTN technique along with other essential planning, search and constraint satisfaction technologies to realise planners that are useful in real applications and at realistic scales.

slide-2
SLIDE 2

2

Practical HTN Planning 2

Literature

  • Human Planning
  • Klein, G. (1998) Sources of Power: How People Make Decisions, MIT Press.
  • Refinement Search
  • Kambhampati, S., Knoblock, C.A. and Yang, Q. (1995) Planning as Refinement Search: A Unified

Framework for Evaluating Design Tradeoffs in Partial-Order Planning, Artificial Intelligence, Vol. 76, No. 1- 2, pp. 167-238, Elsevier.

  • Nonlin
  • http://www.aiai.ed.ac.uk/project/nonlin/
  • Tate, A. (1977) Generating Project Networks, Proceedings of the Fifth International Joint Conference on

Artificial Intelligence (IJCAI-77) pp. 888-893, Boston, Mass. USA, August 1977.

  • O-Plan
  • http://www.aiai.ed.ac.uk/project/oplan/
  • Currie, K and Tate, A. (1991) O-Plan: the Open Planning Architecture, Artificial Intelligence Vol. 52, No. 1,

pp 49-86, Elsevier.

  • Other Practical Planners
  • Ghallab, M., Nau, D. and Traverso, P., Automated Planning – Theory and Practice, chapters 19, 22 and
  • 23. Elsevier/Morgan Kaufmann, 2004.

Literature

Further practical planning systems are described in several chapters of the course textbook… e.g. chapters 19, 22 and 23. Malik Ghallab, Dana Nau, and Paolo Traverso. Automated Planning – Theory and Practice. Elsevier/Morgan Kaufmann, 2004.

slide-3
SLIDE 3

3

Practical HTN Planning 3

Overview

Human Approaches to Planning Practical HTN Planning Refinement Planning as a Unifying View Nonlin and O-Plan Features QA (Modal Truth Criterion) Time, Resource and Other Constraint Handling I-X/I-Plan Overview

slide-4
SLIDE 4

4

Practical HTN Planning 4

  • Expansion of a high level abstract plan into greater detail

where necessary.

  • High level ‘chunks’ of procedural knowledge (Standard

Operating Procedures, Best Practice Processes, Tactics Techniques and Procedures, etc.) at a human scale - typically 5-8 actions - can be manipulated within the system.

  • Ability to establish that a feasible plan exists, perhaps for a

range of assumptions about the situation, while retaining a high level overview.

  • Analysis of potential interactions as plans are expanded or

developed.

  • Identification of problems, flaws and issues with the plan.
  • Deliberative establishment of a space of alternative options,

perhaps based on different assumptions about the situation involved, of especial use ahead of time, in training and rehearsal, and to those unfamiliar with the situation or utilising novel equipment.

Some Planning Features

slide-5
SLIDE 5

5

Practical HTN Planning 5

  • Monitoring of the execution of events as they are expected to

happen within the plan, watching for deviations that indicate a necessity to re-plan (often ahead of this becoming a serious problem).

  • Represent the dynamic state of the world at points in the plan

and use this for ‘mental simulation’ of the execution of the plan.

  • Pruning of choices according to given requirements or

constraints.

  • Situation dependent option filtering (sometime reducing the

choices normally open to one ‘obvious’ one.

  • Satisficing search to find the first suitable plan that meets the

essential criteria.

  • Heuristic evaluation and prioritisation of multiple possible

choices within the constrained search space.

  • Uniform use of a common plan representation with

embedded rationale to improve plan quality, shared understanding, etc.

More Planning Features

slide-6
SLIDE 6

6

Practical HTN Planning 6

Previous slides describe aspects of problem

solving behaviour observed in expert humans working in unusual or crisis situations.

Gary Klein, “Sources of Power”, MIT Press, 1998. But they also describe the hierarchical and mixed

initiative approach to planning in AI developed

  • ver the last 30 years.

Human Approach

slide-7
SLIDE 7

7

Practical HTN Planning 7

HTN Planning is a useful paradigm… Compose workflows/processes from

requirements and component/template libraries

Covers simple through to very complex (pre-

planned) components

Allows for execution support, reactive repair,

recovery, etc.

Suited to mixed initiative (people and systems)

planning and execution

Gives an understandable framework within

which specialised constraint solvers, domain- specific planners (e.g. route finders), optimisers, plan analysers and simulators can work

HTN - Planning Approach

slide-8
SLIDE 8

8

Practical HTN Planning 8

A1 A2 A3 A5 A4

“Initial” Plan Refine Introduce activities to achieve preconditions Resolve interactions between conditions and effects Handle constraints (e.g. world state, resource, spatial, etc.) “Final” Plan

A2.2 A2.1 A1 A3 A5 A4

Plan Library

A2 Refinement

S2 S1

HTN - Activity Composition

slide-9
SLIDE 9

9

Practical HTN Planning 9 Refine Plan Library

Ax Refinement

S2 S1

P

Initial Plan can be any combination of Activities and Constraints

“Refined” Plan

A1.2 A1.1

Q P

“Initial” Plan

P Q

HTN – Initial Plan as “Goals”

slide-10
SLIDE 10

10

Practical HTN Planning 10

Hierarchical Task Network Planning Partial Order Planner Plan Space Planner Goal structure-based plan development - considers

alternative “approaches” only based on plan rationale

QA/Modal Truth Criterion Condition Achievement Condition “Types” to limit search “Compute Conditions” for links to external data and

systems (attached procedures)

Time and Resource Constraint checks

Nonlin core is basis for text book descriptions of HTN Planning

Nonlin (1974-1977)

slide-11
SLIDE 11

11

Practical HTN Planning 11

Nonlin Domain Language – TF

  • pschema makeon

pattern {on $*x $*y} expansion 1 goal {cleartop $*x} 2 goal {cleartop $*y} 3 action {put $*x on top of $*y}

  • rderings 1 ---> 3 2 ---> 3

vars x undef y undef; end;

  • pschema makeclear

pattern {cleartop $*x} expansion 1 goal {cleartop $*y} 2 action {put $*y on top of $*z}

  • rderings 1 ---> 2

conditions usewhen {on $*y $*x} at 2 usewhen {cleartop $*z} at 2 vars x <:non table:> y undef z <:et <:non $*x:> <:non $*y:> :>; end; actschema puton pattern {put $*x on top of $*y} conditions usewhen {cleartop $*x} at self usewhen {cleartop $*y} at self usewhen {on $*x $*z} at self effects + {on $*x $*y}

  • {cleartop $*y}
  • {on $*x $*z}

+ {cleartop $*z} vars x undef y undef z undef; end; always {cleartop table}; initially {on c a} {on a table} {on b table} {cleartop c} {cleartop b} ; plan goal {on a b} goal {on b c};

$*x is a variable

“typed” condition restricts search space example of search control knowledge

Nonlin Task Formalism (TF) Domain Description language Op schemas are the HTN “Methods”. Act schemas are the primitive operator schemas that are not further decomposable.

slide-12
SLIDE 12

12

Practical HTN Planning 12

QA in a partially ordered network of nodes Way to establish value of a condition P=V at some point in the plan Yes/no/maybe responses Alternative Terminology:

  • Contributors, deletors (Austin Tate, Nonlin, QA, Edinburgh, 1975-7)
  • White nights and clobberers (David Chapman, MIT, MTC, 1987, 1st Formalisation)
  • Producers, consumers (Some textbooks)

Initially just allowed imposition of orderings on nodes for a

condition, a b (ordering)

Later also allowed variables within condition to be constrained –

= (codesignation), ≠ (non-codesignation)

Intuitively, a white knight is an activity which re-establishes a

clobbered precondition p

A clobberer in a plan can be "defeated" by imposing ordering or

codesignation/non-codesignation constraints on the plan, or by inserting a white knight between the clobberer and the point where a condition is needed

QA/Modal Truth Criterion

slide-13
SLIDE 13

13

Practical HTN Planning 13

QA/Modal Truth Criterion

Before After

P=V

Contributor No Effect Deletor

Need to ensure no deletor appears between a chosen contributor and point of need

Note that this lends itself to parallel implementations of the algorithms in the network of nodes. Note also that least commitment approaches can be used by using multiple disjunctive contributors to be kept unless some deletor needs to be excluded from the ranges.

slide-14
SLIDE 14

14

Practical HTN Planning 14

O-Plan (1983-1999) Features

  • Hierarchical Task Network Planning
  • Nonlin-like goal-structure, QA and Typed/Compute conditions
  • Partial-Plan “Refinement “ Approach
  • Plan State has “flaws”/issues attached
  • Agenda Architecture with Plan Modification Operations
  • “Opportunistic Search” (agenda type, branch1/branch N)
  • Multiple constraint managers with yes/no [and maybe] results
  • Least Commitment Approach (on activity ordering,
  • bject/variable bindings and other constraints)
  • Constraint “Posting” rather than explicit commitments

(and/or trees with sets of “before” temporal constraints and variable binding (= and ≠) constraints) [as in MOLGEN]

  • Goal structure recording and monitoring to preserve plan

rationale

slide-15
SLIDE 15

15

Practical HTN Planning 15

O-Plan (1983-1999) Features

slide-16
SLIDE 16

16

Practical HTN Planning 16

O-Plan Domain Language – TF

types objects = (a b c table), movable_objects = (a b c); always {cleartop table}; schema puton; vars ?x = ?{type movable_objects}, ?y = ?{type objects}, ?z = ?{type objects}; vars_relations ?x /= ?y, ?y /= ?z, ?x /= ?z; expands {puton ?x ?y};

  • nly_use_for_effects

{on ?x ?y} = true, {cleartop ?y} = false, {on ?x ?z} = false, {cleartop ?z} = true; conditions only_use_for_query {on ?x ?z} achieve {cleartop ?x} achieve {cleartop ?y}; end_schema;

“typed” condition restricts search space example of search control knowledge

?x is a variable

O-Plan Task Formalism Domain Description language. Single schema for block stacking domain. Note “only_use_for_effects” is for ALL effects, since this is a general schema that can be used in any way. Generally in large domains you restrict the effects you are willing to allow a scheme to be selected to achieve to be the “main effects” it is meant to be used for. The condition type “only_use_for_query” means the schema would not allow this particular condition to ever be ACHIEVED by inserting new activities. This in turn restricts the search space, but means the planner cannot search the complete space of all possible plans in the domain… which seeks to prevent the generation of plans with redundant activities.

slide-17
SLIDE 17

17

Practical HTN Planning 17

O-Plan Agent Architecture

slide-18
SLIDE 18

18

Practical HTN Planning 18

O-Plan Agent Architecture

Later became

  • Issues
  • Nodes
  • Constraints
  • Annotations

Later became Plan Modification Operators

slide-19
SLIDE 19

19

Practical HTN Planning 19

O-Plan Planning Workflow

slide-20
SLIDE 20

20

Practical HTN Planning 20

Human relatable and presentable objectives, issues,

sense-making, advice, multiple options, argumentation, discussions and outline plans for higher levels

Detailed planners, search engines, constraint solvers,

analyzers and simulators act in this framework in an understandable way to provide feasibility checks, detailed constraints and guidance

Sharing of processes and information about process

products between humans and systems

Current status, context and environment sensitivity Links between informal/unstructured planning, more

structured planning and methods for optimisation

A More Collaborative Planning Framew ork

slide-21
SLIDE 21

21

Practical HTN Planning 21 Shared, intelligible, easily communicated and

extendible conceptual model for objectives, processes, standard operating procedures and plans:

  • I

Issues

  • N

Nodes/Activities

  • C

Constraints

  • A

Annotations

Communication of dynamic status and presence for

agents, and reports about their collaborative processes and process products

Context sensitive presentation of options for action Intelligent activity planning, execution, monitoring, re-

planning and plan repair via I-Plan and I-P2 (I-X Process Panels)

I-X/I-Plan (2000- )

slide-22
SLIDE 22

22

I – Issues - Plan Agenda - implied constraints on the plan N – Nodes - Plan Entities - Anchoring activities in a plan C – Detailed Constraints - Detailed Plan Constraints A – Annotations A model of a product being created, modified or used within an I-X system or agent is specified by giving a set of constraints on the whole space of legitimate products in the area of application. Nodes are the principal entity in the

  • product. The decision to include a node is a very important

step in a system whose aim is to synthesise such a product description or model.

Practical HTN Planning 22

Constraints Issues Nodes

Plan State

Space of Legitimate Behaviours

Issues or Implied Constraints Node Constraints Detailed Constraints I N C A Annotations

<I-N-C-A> Framew ork

slide-23
SLIDE 23

23

The I-X processing cycle begins with the selection of one or more of the outstanding issues in a product model. A handler is selected to address the issue(s). The handler is then called. This can lead to alterations to:

  • 1. the nodes included in the product model (a key decision

for synthesis systems);

  • 2. the introduction of constraints on the way in which the

nodes are related between themselves and to other entities in the application domain which are termed Node-Relatable Objects

  • 3. the introduction of further issues or sub-issues to

handle. A constraint propagation loop is then activated which looks at the feasibility of introducing the constraints added by the issue handling loop. This can lead to further constraints being introduced to resolve constraint interactions, and can lead to even further issues which are passed back to the Issue Handler – possibly just to be posted into the Product Model for later handling.

Practical HTN Planning 23

Constraints Issues Nodes

Plan State

Space of Legitimate Behaviours

Issues or Implied Constraints Node Constraints Detailed Constraints I N C A Annotations

<I-N-C-A> & I-X

Do (IH) Choose (IH)

IH=Issue Handler (Agent Functional Capability)

Propagate Constraints

slide-24
SLIDE 24

24

Practical HTN Planning 24

Anatomy of an I-X Process Panel

slide-25
SLIDE 25

25

Practical HTN Planning 25

I-P2 aim is a Planning, Workflow and Task Messaging “Catch All”

  • Can take ANY requirement to:
  • Handle an issue
  • Perform an activity
  • Respect a constraint
  • Note an annotation
  • Deals with these via:
  • Manual activity
  • Internal capabilities
  • External capabilities
  • Reroute or delegate to other panels or agents
  • Plan and execute a composite of these capabilities (I-Plan)
  • Receives reports and interprets them to:
  • Understand current status of issues, activities and constraints
  • Understand current world state, especially status of process products
  • Help user control the situation
  • Copes with partial knowledge of processes and organisations
slide-26
SLIDE 26

26

Process Panel

I-X Process Panel and Tools

Domain Editor Messenger I-Plan Map Tool

slide-27
SLIDE 27

27

I-X for Emergency Response

Collaboration and Communication

Command Centre

Central Authorities Isolated Personnel

Emergency Responders

slide-28
SLIDE 28

28

Planning Research Areas & Techniques Planning Research Areas & Techniques

Problem is to make sense

  • f all these techniques

Problem is to make sense

  • f all these techniques

Deals with whole life cycle of plans

  • Search Methods

Heuristics, A*

  • Graph Planning Algthms GraphPlan
  • Partial-Order Planning

Nonlin, UCPOP

  • Hierarchical Planning

NOAH, Nonlin, O-Plan

  • Refinement Planning

Kambhampati

  • Opportunistic Search

OPM

  • Constraint Satisfaction

CSP, OR, TMMS

  • Optimisation Methods

NN, GA, Ant Colony Opt.

  • Issue/Flaw Handling

O-Plan

  • Plan Analysis

NOAH, Critics

  • Plan Simulation

QinetiQ

  • Plan Qualitative Mdling

Excalibur

  • Plan Repair

O-Plan

  • Re-planning

O-Plan

  • Plan Monitoring

O-Plan, IPEM

  • Plan Generalisation

Macrops, EBL

  • Case-Based Planning

CHEF, PRODIGY

  • Plan Learning

SOAR, PRODIGY

  • Plan Generalisation

Macrops, EBL

  • Case-Based Planning

CHEF, PRODIGY

  • Plan Learning

SOAR, PRODIGY

  • User Interfaces

SIPE, O-Plan

  • Plan Advice

SRI/Myers

  • Mixed-Initiative PlanS

TRIPS/TRAINS

  • Planning Web Services

O-Plan, SHOP2

  • Plan Sharing & Comms

I-X, <I-N-C-A>

  • NL Generation

…

  • Dialogue Management

…

  • Domain Modelling

HTN, SIPE

  • Domain Description

PDDL, NIST PSL

  • Domain Analysis

TIMS

slide-29
SLIDE 29

29

Practical HTN Planning 29

Summary

Human Approaches to Planning Practical HTN Planning Refinement Planning as a Unifying View Nonlin and O-Plan Features QA (Modal Truth Criterion) Time, Resource and Other Constraint Handling I-X/I-Plan Overview