Integrating Agents, Ontologies, and Semantic Web Services for Collaboration on the Semantic Web

Integrating Agents, Ontologies, and Semantic Web Services
for Collaboration on the Semantic Web

Michael Stollberg and Thomas Strang

DERI

–

Digital Enterprise Research Institute


First International Symposium on Agents and the Semantic Web

AAAI Fall Symposium Series 2005

Arlington, Virginia, 05 November 2005


2

Content

1.
Motivation and Aims

–
Semantic Web vision = automated collaboration support on the Web

–
Aim = technical realization


2.
Framework

–
Conceptual Model

–
Elements (Description and Usage)

•
Model of Agency

•
“Collaboration Services”

•
Goal
-
driven Architecture


3.
Semantic Collaboration Management

–
Collaboration Establishment

–
Collaboration Execution


4.
Conclusions


3

Motivating Example

L

M

P

D

Agent A(P)

Goal:

plan[


-

med. treatment for M

-

chauffeur(P)

-

consistent with calendar(P)]

Plan:


-

check plan from A(L)

-

if OK, then book


else: revise plan

Agent A(L)

Goal:

plan[


-

med. treatment for M

-

chauffeur(P)]

Plan:

1) find suitable doctor

2) coordinate chauffeur(P)

med. info
service

calendar

service

transport

info serv.

owns

owns

O
-
person

O
-
medical

O
-
date&time

O
-
transport

Agent A(D)

Goal:

offer[


-

med. treatment

-

appointement]

Plan:

1) find customer

2) book appointement

owns

book app.
service

doctor

information

(website)

ontology usage

(also for agents,

omitted here)

1:need medical treatment

2: inform, chaffeur(P)

3a: assign goal

3b: make /

find plan

3b: make /

find plan

3a: assign goal

Xb: make /

find plan

Xa: assign goal

4:find
doctor

uses

provides

5:inform, plan

6:revise plan

7:book appointment

interact

uses

8: add plan

from: Berners
-
Lee, T.; Hendler, J.; Lassila, O.:
The Semantic Web. A new form of Web Content that is meaningful

to computers will unleash a revolution of new possibilities.
In: Scientific American 9(5), May 2001.

4

Automated Collaboration Support on
the Semantic Web

•
In fact, the Semantic Web vision proclaims
seamless and automated
collaboration support
on the Web

–
collaboration = cooperative interactions of individuals for achieving complex
objectives or objectives that require purposeful cooperation

–
... is a constituting principle of (human) society

–
… the “Semantic Web” shall provide sophisticated support for this


•
Base Technologies:

–
Software Agents


personal assistants / electronic representatives

–
Ontologies


semantic interoperability

–
(Semantic) Web Services


computation over the Web


•
Technology for Collaboration Support needs to:

–
reflect the epistemology of collaboration

–
integrate base technologies properly

–
comply with technical requirements



5

Our Approach

•
Agents as electronic representatives of collaborating real
-
world entities use:

–
(Semantic) Web resources as information resources

–
(Semantic) Web services as means for automated communication & cooperation

•
Semantic Web Service technologies for collaboration management

•
Semantic descriptions as extension of WSMO


Design Principles

(combining Agent & Web technology for collaboration support)

1.
Agents as Symmetric Collaboration Entities

2.
Ontologies as Data Model

3.
Web Compliance

4.
Goal Driven Architecture

5.
Semantic Collaboration Management

6.
Strict Decoupling and Strong Mediation

7.
Maximal Automation

6

Conceptual Model

Agent B

Goal

Instance

Agent A

Goal
Instance

has(1,n)

compatible goals =

collaboration partners


automated collaboration

execution

Domain Knowledge

uses(1,n)

has(1,n)

uses(1,n)

Collaboration

Service

Web

Service

Goal Template

Repository

Service

Repository

Collaboration

Service

Web

Service

Owner

Owner

goal
assignment

service
discovery

owns

Ontology

Ontology

Ontology

Ontology

7

Model of Agency

•
Agent = electronic representative of a real world entity
involved in collaborative interactions

–
receives task from owner (objectives represented as Goals)

–
uses “Collaboration Service” as facility for participating in collaborations
automatically executed over the Semantic Web

–
controls own collaboration behavior by using services for collaboration
management

•
Goal
-
driven Collaborative Interface Agents


–
Reactivity & Proactiveness

via interaction with owner and other agents

–
Autonomy
as self
-
controlled collaboration behavior

–
Social Abilities

as purposeful information interchange & communication via
Collaboration Services

•
Semantic description

–
‘data container’ of owner and collaboration information

–
ontologies used as data model

8

Agent Semantic Description

Agents

electronic representative of real world entity that wants to achieve

an objective by collaborative interaction with other
real world entities

Class

agent


hasNonFunctionalProperties
type

nonFunctionalProperties


importsOntology
type

ontology


usesMediator
type

ooMediator


owner
type

owner


collaboration
type

collaboration


history
type

collaboration

Class

owner


hasNonFunctionalProperties
type

nonFunctionalProperties


owner
type

instance


preference
type

axiom


policy
type

axiom


serviceUsagePermission
type

service

Class

collaboration


hasNonFunctionalProperties
type

nonFunctionalProperties


goal
type

goal
single
-
valued


service
type

service

9

Collaboration Services

•
Web Service used by Agents to participate in collaborations
automatically executed over the (Semantic) Web

–
a computational facility accessible over the Web that provides a functionality
useful for an agent to automate collaboration execution

–
can provide very simply or arbitrary complex functionality

–
can “talk” to other agents via their collaboration services

•
dynamically
detected

(possibly
composed
) and
consumed

by
an Agent wrt / via an Goal Instance

–
each objective is solved in a different collaboration

–
agent discovers suitable Collaboration Services (Web service discovery),
possibly determines composite Collaboration Service

–
Goal Instance +Agent data container = input for service execution

•
Semantic description:

–
extended WSMO Web service description

–
“orchestration” defines interaction behavior with other services / resources
for achieving functionality

10

Collaboration Service Description & Usage


Agent A

postcondition

Goal Instance

effect

submission

result


Service

Implementation

(not of interest for

service description)







Capability

Client
-
Interface

WS

Orchestration


Agent B


Service

Implementation

(not of interest for

service description)






WS

Orchestration

Client
-
Interface

(1) = functional suitability

(2) = service usage

(3) = service interaction

(1)

(2)

(1)

(2)

(3)

preferences

postcondition

Goal Instance

effect

submission

result

preferences

WS

nFP

Capability

nFP

11

Collaboration Service Description

Collaboration Services

facility an agent uses for participating in an automatically executed collaboration;

interacts with other agents and utilizes Semantic Web resources via its orchestration

Class

collaborationService
is
-
a
wsmoService


hasNonFunctionalProperties
type

nonFunctionalProperties


importsOntology
type

ontology


usesMediator
type

ooMediator


hasCapability
type

capability
single
-
valued


hasSharedVariables
type

sharedVariable


hasPrecondition
type

axiom


hasAssumption
type

axiom


hasPostcondition
type

axiom


hasEffect
type

axiom


hasInterface
type

interface


clientInterface
type

serviceInterfaceDescription


orchestration
type

serviceInterfaceDescription

Class

serviceInterfaceDescription
sub
-
Class
wsmoServiceInterface


hasNonFunctionalProperties
type

nonFunctionalProperties


importsOntology
type

ontology


usesMediator
type

ooMediator


hasVocabulary
type

concept_in, concept_out, concept_controlled


hasState
type

ontology


hasGuardedTransition
type

if (condition) then
rule


// rule = if
-
then rule | add
-
delete
-
update(fact) | forAll / choose (var)
rule



12

Goals

•
Goal
-
driven Architecture

–
users specify objectives as Goal Instances and assign this to their agent

–
goal creation by instantiation of Goal Templates

–
each objective is solved in a different collaboration

> Goals are the central element of the system

•
Goal = semantic description of an user’s objective that carries
all information needed for automated collaboration execution
via Collaboration Services

–
user objective as “final desired state”

–
constraints on goal resolution process

–
preferences / policies on: collaboration partners, services, time

–
Collaboration Service usage by Goal Instance + Agent data container

•
semantic description:

–
extended WSMO Goal description

–
Management by Goal Templates, Goal Instances, Goal Ontologies

13

Goal Templates and Instances

Goal Template


-
schematic goal description

o
schema of desired final state

o
constraints on resolution process (integrity constraints on states)

-
pre
-
defined at “design time”, maintained by “ administrator


Goal Instance


-
concrete objective specification as Goal Template instantiation

o
proper semantic refinement of its Goal Template wrt. final desired state

o
additional information for service usage

-
created at runtime by users (agent owners) or other agents


after resolution, a Goal Instance can become a Goal Template


14

Goals Semantic Description

Goal Template

predefined schema of user objective (WSMO 1.0 goal plus constraints)

Class

goalTemplate
subClassOf

wsmov1.0Goal


hasNonFunctionalProperties
type

nonFunctionalProperties


importsOntology
type

ontology


usesMediator
type

{ooMediator, ggMediator}


hasPostcondition
type

axiom


hasEffect
type

axiom


hasConstraint
type

axiom


Goal Instance

concrete objective assigned to an agent (instantiated Goal Template, client for service usage)

Class

goalInstance
sub
-
Instance

goalTemplate


hasNonFunctionalProperties
type

nonFunctionalProperties


hasPostcondition
type

axiom


hasEffect
type

axiom


hasConstraint
type

axiom


hasSubmission
type

instance


hasResult
type

instance


goalResolutionStatus
type

goalResolutionStatus


15

Goals Ontologies

-
Goal Ontologies

for efficient management:

o
Nodes = Goals, Arcs = ggMediators that denote “functional correspondence”

o
computed by
Δ

= explicit logical difference between final desired states


GI

{S
1
,S
2
,S
3
}

{S
2
,S
3
}

GT
-
1

Book Flight

GT
-
2

Book Flight

(Star Alliance)

subsume

{S
3
}

GT
-
3

Book Hotel

subsume

{S
5
,S
8
}

{
S
2
,S
3
,
S
5
, S
8
}

GT
-
4

Book US
-
Trip

(flight + hotel)

subsume

GI

{(S
3
, S
8
)}

{(
S
3
,
S5),(
S
3
,
S
8
)}

composition

16

Collaboration Management

Partner
Discoverer

Service
Discoverer

Choreography
Discoverer

GI

{GI}

GI

{S}


boolean

{S}

Collaboration
(preliminary)

(A
i
(G
1
, {S}
1
),



A
2

(G
2
, {S}
2
), ..)

Goal

Instance

Agent

(A
i
(G
1
, {S}
1
-
reduced
),



A
2

(G
2
, {S}
2
-
reduced
),..)


Goal

Instance

…

separately for each GI

each possible service combination

Agent

Collaboration
(final)

Collaboration
Executor

each collaboration

goal resolution for

the goal of each

collaboration partner

goal resolution for

the goal of each

collaboration partner

controlled individually by each agent, functional components as Web Services

17

Collaboration Management Components

Collaboration Establishment

applying “conventional” Semantic Web Service techniques


•
Partner Discovery = find agent with compatible goal

–
compatible final states of Goal Instances of agents

•
Service Discovery = find usable Collaboration Service

–
facility for participating in automated collaboration execution

–
goal final state <> service post
-
state ^ submission <> service pre
-
state

–
composition as sub
-
task of service discovery

•
Choreography Discovery = valid Collaboration Service interaction

–
orchestration compatibility of Collaboration Service of collaborating agents

–
existence of a valid interaction protocol for all needed interactions

Collaboration Execution

execute collaborations as Web Service interaction:

–
feasibility determination (availability of resources, duplication handling)

–
monitors & controls execution

18

Partner Discoverer Architecture

GI
i

Action
-
Resource
Ontology

DiscoveryResult

sets of compatible Goal Instances

(2) GG
Matcher

Discovery Request

initiating Goal Instance

GT
i

(1) Cooperation
Knowledge Filter

GT
g

GI
g

GI
g

instanceOf

instanceOf,

status = ‘open’


Action Compatibility

19

Service Discoverer Architecture

Discovery Request

Goal Instance

Discovery Result

usable Services

Service
Repository

Discovery Result

(intermediary)

Service

Filter

(2) GIS Matcher

GI
i

GT
i

instanceOf

(1) Pre
-
Selector

Action

Equality

20

Choreography Discoverer Architecture

Discovery Request

set of services {S}

Discovery Result

Boolean (service compatibility)

(2) SC
comm

Checker

Choreography Description Translator


(1) SC
info

Checker

WSDL2StateSign

BPEL2ASM

yes

no

21

Conclusions

•
Framework for Collaboration on the Semantic Web

–
agents, goals, Web Services “integrated” for collaboration support

–
SWS technologies for

–
prototype implementation in Semantic Web Fred (SWF)


•
Important Aspects

–
Model of Agency

–
Collaboration Services

–
Goal
-
driven Architecture

–
Semantic Collaboration Management


•
Lessons Learned

–
core (reasoning) tasks are the same in agents and SW / SWS

–
rationale agent paradigm = determine best next action for achieving goal

–
(Semantic) Web Service = determine complete interaction model before
execution

=> which one is “better”, i.e. more applicable


22



< Q&A>



< Thank You />

</ The Big Dream>