Agent Orientation
and
Information Systems
Eric Yu
University of Toronto
Presentation at Tsinghua University, Beijing, China
July 8, 1999
Information Systems research at the
University of Toronto
Dept. of Computer Science
 Databases, Information Systems
 Artificial Intelligence, Knowledge Rep.
 Software Engineering, Requirements Eng.
Dept. of Industrial Engineering
Faculty of Management
Faculty of Information Studies
Knowledge Media Design Institute
2
@ Eric Yu 1999
Outline of this talk
1. An Emerging Paradigm in Computing
2. Agent Orientation for Enterprise Information Systems?
3. An Agent-Oriented Modelling Framework
i*
4. Research Directions
3
@ Eric Yu 1999
AOIS workshop
@ AutonomousAgents’99 Seattle USA May 1, 1999
@ CAiSE*99 Heidelberg Germany June 14-15, 1999
Invited speakers
 Katia Sycara (Carnegie-Mellon Univ.)
 Mike Huhns (Univ. S.Carolina)
 John Mylopoulos (Univ. Toronto)
 Cristiano Castelfranchi (Psych., NRC, Italy)
 Stefan Kirn (TU-Ilmenau, Germany)
2 panels, 9 contributed papers
http://aois.org
4
@ Eric Yu 1999
Part 1
Agent-Orientation as an emerging
paradigm in Computing
Programming Paradigms
Agent Abstractions
Agent Architectures
5
@ Eric Yu 1999
Programming Paradigms
1950’s -- Machine and assembly language
1960’s -- Procedural programming
1970’s -- Structured programming
1980’s -- Object-based and declarative programming
1990’s -- Frameworks, design patterns, scenarios, protocols, and
components (ActiveX/COM and Java Beans)
The trend has been from elements that represent abstract
computations to elements that represent the real world
[Huhns AOIS’99]
6
Features of
Languages and Paradigms
Concept
Procedural Language
Object Language
Multiagent Language
Abstraction
Building Block
Computation model
Design Paradigm
Architecture
Type
Instance, Data
Procedure/Call
Tree of procedures
Functional decomposition
Modes of Behavior
Terminology
Coding
Implement
Class
Object
Method/Message
Interaction patterns
Inheritance and
Polymorphism
Designing and using
Engineer
Society, Team
Agent
Perceive/Reason/Act
Cooperative interaction
Managers, Assistants,
and Peers
Enabling and enacting
Activate
[Huhns AOIS’99]
7
Agent Abstractions
Agent abstractions are mentalistic
• beliefs: agent’s representation of the world
• knowledge: (usually) true beliefs
• desires: preferred states of the world
• goals: consistent desires
• intentions: goals adopted for action
Multi-agent abstractions involve interactions
• social: about collections of agents
• organizational: about teams and groups
• ethical: about right and wrong actions
• legal: about contracts and compliance
[Huhns AOIS’99]
8
@ Eric Yu 1999
Why Do These Abstractions Matter?
Because modern applications go beyond
traditional metaphors and models in terms of their
dynamism, openness, and trustworthiness
 virtual enterprises: manufacturing supply
chains, autonomous logistics
 electronic commerce: utility management
 communityware: social user interfaces
 problem-solving by teams
[Huhns AOIS’99]
9
@ Eric Yu 1999
Agent architectures
Deliberative
Agents
Reactive
Agents
Hybrid
Agents
Agent
Architectures
Interacting
Agents
Other
Approaches
[Kirn AOIS’99]
10
@ Eric Yu 1999
Reactive Agents
World
Controller
Stimuli
S
e
n
s
o
r
Plans
Pattern 1
Plan 1
Pattern 2
Plan 2
.
.
.
.
.
.
Pattern n
Plan n
E
f
f
e
c
t
o
r
Agent
[Kirn AOIS’99]
11
@ Eric Yu 1999
Deliberative Agents
World
Cognition
Inference
Strategies
Memory
S
e
n
s
o
r
Goals
Environment
Model
Utility
Function
Interpretation
Domain
Knowledge
Planner
E
f
f
e
c
t
o
r
Agent
[Kirn AOIS’99]
12
@ Eric Yu 1999
Types of Information Agents
“Standard”
information agents
and architectures are
becoming available
Application
Program
User Interface
Agent
Reg/Unreg
(KQML)
Broker
Agent
Reply
Reply
Query or
Update
In SQL
Ontology Agent
Reg/Unreg (KQML)
Reg/Unreg
(KQML)
Mediator
Agent
Ontology
(CLIPS)
Mediated
Query (SQL)
Reg/Unreg
(KQML)
Mediated
Query (SQL)
Reply
Schemas
(CLIPS)
Database Resource Agent
Reply
Database Resource Agent
[Huhns AOIS’99]
13
@ Eric Yu 1999
Part 2
Agent-Orientation for Enterprise
Information Systems
The Changing Nature of Enterprise
The Challenge for Enterprise Systems
Why Agent-Orientation for Enterprise Systems
14
@ Eric Yu 1999
The Changing Nature of Enterprise
distributed and networked
 people, organization, and work practices, not
just the technology!
diversity, local autonomy, open-endedness
 much uncertainty, incomplete knowledge &
control
 need flexibility
change and evolution
 constantly and rapid
15
@ Eric Yu 1999
The Challenge for Enterprise Systems
need to deal with conflicting needs and demands
from many players / stakeholders
From Integration to Cooperation
Autonomous
Islands
16
Cooperation
“working together”
Full
Integration
@ Eric Yu 1999
Why Agent-Orientation for
Enterprise Information Systems
Agent orientation addresses the demands and
challenges of new enterprise environments and
systems
What would it mean?
We should develop Agent-Oriented...
 requirements engineering techniques, models
 design and architectural approaches
 implementation methods and technologies
 run-time and evolution support
17
@ Eric Yu 1999
Part 3
An Agent-Oriented Modelling
Framework i*
Understanding “Why” — intentionality
Strategic Dependencies
Strategic Rationales
Analysis and Design Support
Knowledge Representation
18
@ Eric Yu 1999
Modelling for Enterprise Systems
It is well-recognized that many types of
modelling are required to deal with the various
aspects of enterprise, e.g.,
 activity modelling
 function modelling
 resource modelling
 information modelling
 organization modelling
e.g., CIMOSA, GERAM,...
19
@ Eric Yu 1999
Towards richer organization
modelling
How do we express and reason about
 motivations
 rewards
 different ways for achieving objectives
• understanding “why”
• opportunities and vulnerabilities
… strategic business and social relationships
20
@ Eric Yu 1999
Consider one very successful
enterprise...
important organizational and social
aspects are missing in conventional models
21
@ Eric Yu 1999
Modelling Strategic Actor
Relationships and Rationales
- the i* modelling framework

 have goals, beliefs, abilities, commitments
 depend on each other for goals to be achieved,
tasks to be performed, resources to be furnished
 are semi-autonomous -- not fully knowable /
controllable
22
@ Eric Yu 1999
Wants and Abilities
I want
...
23
I can
provide
...
@ Eric Yu 1999
A Strategic Dependency Model
LEGEND
actor
24
goal dependency
task dependency
resource dependency
softgoal dependency
@ Eric Yu 1999
Roles, Positions, Agents
LEGEND
agent
position
role
A Strategic Dependency
model showing reward
structure for improving
performance, based on an
example in [Majchrzak96]
25
@ Eric Yu 1999
Some strategic dependencies between
IKEA and its customers
26
@ Eric Yu 1999
A Strategic Rationale Model
27
@ Eric Yu 1999
Analysis and Design Support
opportunities and vulnerabilities
 ability, workability, viability, believability
 insurance, assurance, enforceability
 node and loop analysis [Yu ICEIMT’97]
design issues
 raising, evaluating, justifying, settling
 based on qualitative reasoning
[Chung Nixon Yu Mylopoulos, forthcoming monograph]
28
@ Eric Yu 1999
Means-Ends Analysis
Claims
Handling
Actor
boundary
Get accident
info
Handle
claim
Verify
policy
Settle
claim
Prepare
offer
Determine
fault
Whose
fault?
Settlement
cost?
Determine
cost to settle
D
Witness
29
[Mylopoulos AOIS’99]
D
Doctor
D
Injury
info
D
Sufficient
treatment
D
Police
Accident
info
Minimal
repairs
Appraise
damage
Appraiser
@ Eric Yu 1999
Sample i* representation for an actor (in Telos)
TELL Class Physician IN PositionClass
ISA ProfessionalPosition WITH
resDepends, committedTo
fs: FeeForTreatment WITH
dependee
cm:ClaimsManager end
goalDepended, commitsTo
td: $Treated(p.injury)$ WITH
depender
p:Patient end
taskDepends, committedTo
tm: TakeMedication(p.med) WITH
dependee
p:Patient end
covers
tp: TreatingPatient(p)
bi: Billing(p.insurCo)
integrityConstraint
correctClaimsManager:
$cm=p.insurCo.claimsMgr$
end
30
@ Eric Yu 1999
The Strategic Rationale Model
- a partial schema
Title:
Creator:
idraw
Prev iew :
This EPS picture w as not s av ed
w ith a preview inc luded in it.
Comment:
This EPS picture w ill print to a
Pos tSc ript printer, but not to
other ty pes of printers.
31
@ Eric Yu 1999
Contributions
new ontology
new types of reasoning
applied to business process modelling, enterprise
modelling, requirements engineering, software
process, organization analysis
Some applications by external groups
 software maintenance domain [Briand95, 97]
 CIM [Petit98]
32
@ Eric Yu 1999
Ongoing Work
formal knowledge representation using a
conceptual modelling language Telos
tool building - GUI, repository support
knowledge libraries
• strategic knowhow
• case-based reasoning
• patterns
case studies
coordination with other modelling techniques
33
@ Eric Yu 1999
Related Work
Goal-Oriented and Agent-Oriented Requirements
Engineering (e.g., Feather87, Dardenne93,
Chung93, Bubenko93, Dubois94, Anton97)
see also ISRE, ICRE, RE j., REFSQ.
CSCW, groupware, workflow, COOCS (now
GROUP)
Enterprise Integration (e.g., ICEIMT)
AI, Distributed AI
Organization Theories
34
@ Eric Yu 1999
Part 4
Research Directions
#1 Requirements-Driven Information Systems
Engineering (J. Mylopoulos, E. Yu)
#2 Cooperative Information Systems
(G. DeMichelis, E. Dubois, M. Jarke, F. Matthes, J. Mylopoulos,
M. Papazoglou, J. Schmidt, C. Woo, E. Yu)
#3 Intentionality Management
35
(E. Yu)
@ Eric Yu 1999
Research Direction #1
Requirements-Driven Information
System Engineering
Traditionally, IS Engineering has been implementation-driven.
This means that the programming paradigm of the day dictated the
design and requirements paradigms.
So, structured programming led to structured design and
(requirements) analysis, while object-oriented programming led to
object-oriented design and analysis.
What would requirements-driven
IS Engineering look like??
[Mylopoulos AOIS’99]
36
@ Eric Yu 1999
Why Requirements-Driven?
Requirements analysis is arguably the most important phase of
information system development; that’s where the most and the
costliest errors are introduced in software systems.
The importance of detailed design and implementation will wear off
over time, thanks to software reuse, COTS and the like; requirements
analysis will always be there and will always be important.
Requirements analysis is the phase where technology meets the real
world, where technical considerations have to be balanced against
personal, organizational and social ones; this calls for special skills on
the part of the requirements engineer, and makes the phase particularly
challenging.
[Mylopoulos AOIS’99]
37
@ Eric Yu 1999
Where Are We??
Agent-oriented
programming
i*
KAOS
Z
UML
[Mylopoulos AOIS’99]
38
@ Eric Yu 1999
Where Do We Want To Be??
i*
Agent-oriented
programming
Guiding Principle: Push concepts as far down as
possible (…and see what happens!)
[Mylopoulos AOIS’99]
39
@ Eric Yu 1999
Research Direction #2
Cooperative Information Systems
“Cooperative Information Systems: A Manifesto”
G. DeMichelis, E. Dubois, M. Jarke, F. Matthes, J. Mylopoulos, M.
Papazoglou, K. Pohl, J. Schmidt, C. Woo, E. Yu
in Cooperative Information Systems: Trends and
Directions, M. Papazoglou and G. Schlageter (eds).
Academic Press, 1997.
“A three-faceted view of information systems”
G. DeMichelis, E. Dubois, M. Jarke, F. Matthes, J. Mylopoulos,
J. Schmidt, C. Woo, E. Yu. Communications of the ACM,
December 1998.
40
@ Eric Yu 1999
When is an IS cooperative?
“An information system is cooperative if it shares
goals with other agents in its environment, such as
other information systems, human agents and the
organization itself, and contributes positively
towards the fulfillment of these common goals.”
41
@ Eric Yu 1999
Group
Collaboration
Facet
Organizational
Facet
Systems
Facet
Three Facets of
“Working Together”
42
@ Eric Yu 1999
Complementarity of...
Organizational Facet and
Group Collaboration Facet
defines what is possible
Execution
Group
Col l aborati on
S upport
Organi zati onal
Model
Model
modifies
43
@ Eric Yu 1999
Towards a generic architecture
for Cooperative Information Systems
Organizational
Facet
Group
Collaboration
Facet
Cooperation
Support Layer
Systems Facet
Base Layer
base layer system
systems coop. support agent
organizational coop. support agent
group coop. support agent
multi-facet coop. support agent
repository
44
@ Eric Yu 1999
Research Direction #3
Intentionality Management
Beyond information management
 managing the networks of intentional attitudes
and relationships – goals, beliefs, wants,
abilities, commitments, …
 managing choice, decision making, uncertainty,
openness and freedom
 is an elaboration on an important aspect of
“knowledge management”
45
@ Eric Yu 1999
Intentionality Management (cont’d)
Software development is but one example of
“intentionality-intensive” work environments.
 other examples: enterprise management, virtual
enterprise, product design/development,
“knowledge work” in general, ...
46
@ Eric Yu 1999
Summary and Conclusions
We are at the threshold of a new era for computing.
Agent-Orientation will give us
 (not only) more powerful computing technologies
 (but also) more effective computing that will
better meet enterprise and human needs
• through use of social organization paradigm for
computing
• and incorporation of social organizational
analysis into overall framework for modelling
analysis, and design
47
@ Eric Yu 1999
Sponsors and Partners
Natural Sciences and Engineering Research Council
of Canada (NSERC)
Communications and Information Technology
Ontario (CITO)
Institute for Robotics and Intelligent Systems (IRIS)
Mitel Corporation
IBM
For further information and references, please see
http://www.fis.utoronto.ca/~yu
48
@ Eric Yu 1999