SPIIRAS

CKM Workshop (MIT, Cambridge, MA; January 24, 2006

Context-Driven Decision Making

in Network-Centric Operations:
Agent-Based Intelligent Support

Prof. Alexander Smirnov

Head of Computer Aided Integrated Systems Laboratory
St.Petersburg Institute for Informatics and Automation of the Russian
Academy of Sciences (SPIIRAS)

39, 14th line, St.Petersburg, 199178, Russia

Phone: +7(812) 328-2073; Fax: +7(812) 328-0685;
e-mail: [email protected]
Form Approved
Report Documentation Page OMB No. 0704-0188

Public reporting burden for the collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and
maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information,
including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington
VA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to a penalty for failing to comply with a collection of information if it
does not display a currently valid OMB control number.

1. REPORT DATE 3. DATES COVERED

2. REPORT TYPE
JAN 2006 00-00-2006 to 00-00-2006
4. TITLE AND SUBTITLE 5a. CONTRACT NUMBER
Context-Driven Decision Making in Network-Centric Operations: 5b. GRANT NUMBER
Agent-Based Intelligent Support
5c. PROGRAM ELEMENT NUMBER

6. AUTHOR(S) 5d. PROJECT NUMBER

5e. TASK NUMBER

5f. WORK UNIT NUMBER

7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING ORGANIZATION

REPORT NUMBER
St.Petersburg Institute for Informatics and Automation of the Russian
,Academy of Sciences (SPIIRAS),39, 14th line, St.Petersburg, 199178,
Russia,
9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSOR/MONITOR’S ACRONYM(S)

11. SPONSOR/MONITOR’S REPORT

NUMBER(S)

12. DISTRIBUTION/AVAILABILITY STATEMENT

Approved for public release; distribution unlimited
13. SUPPLEMENTARY NOTES
Collaboration and Knowledge Management (CKM) Workshop, 24-26 Jan 2006, Cambridge, MA
14. ABSTRACT

15. SUBJECT TERMS

16. SECURITY CLASSIFICATION OF: 17. LIMITATION OF 18. NUMBER 19a. NAME OF
ABSTRACT OF PAGES RESPONSIBLE PERSON
a. REPORT b. ABSTRACT c. THIS PAGE Same as 80
unclassified unclassified unclassified Report (SAR)

Standard Form 298 (Rev. 8-98)

Prescribed by ANSI Std Z39-18
SPIIRAS

Presentation Outline

z Introduction
z Knowledge Logistics
z Context-Driven Methodology of Operational Decision Making
z System “KSNet” Research Prototype
z Case Study
z Conclusion & Future Work

2
SPIIRAS

Location of St.Petersburg

3
SPIIRAS

Russian Academy of Sciences

z Founded in 1724
z The research umbrella organization of the Russian Government
z Members of the Academy: Academicians – 458; Corresponding
Members - 686
z 363 units (Research Institutes and Centers)
z 116,500 personnel: 55,100 Researchers (10,000 D.Sc., and
26,000 Ph.D.)

4
STRUCTURE OF THE RUSSIAN ACADEMY OF SCIENCES
SPIIRAS

PREZIDIUM
Yu. S. Osipov, President

Departments representing Regional

Regional Branches
scientific fields Scientific Centres
Far East Siberian Ural
Mathematical Sciences St.
St.Petersburg
Petersburg
G.I.
G.I.Alferov
Alferov
Vice-President
Vice-Presidentofof
RAS
RAS
Physical Sciences Scientific Centres

Information Technologies Amur Buryat Arkhangelsk

Chernogolovka
& Computer Systems
Kamchatka Irkutsk Komi Daghestan

Power & Mechanical Engineering, North- Kemerovo Orenburg Kabardino-

Balkariya
Mechanics & Control Processes
East Krasnoyarsk Perm Kazan
Khabarovsk Novosibirsk Udmurt Karelian
Biological Sciences
Kola
Primorski Omsk Chelyabinsk
Chemistry & Materials Science Puschino
Sakhalim Tomsk Samara

Tyumen Saratov
Social Sciences
Troitsk
Yakutsk
Ufa
Historical & Philological Sciences
Vladikavkaz
South
Earth Sciences

5
SPIIRAS

SPIIRAS

z St.Petersburg Institute for Informatics and Automation (SPIIRAS)

z Founded in 1978
z Only 1 Russian Academy of Science Institute operating in Northwest
Russia in Computer Science discipline
z 210 Personnel: 167 Researchers (34 D.Sc., and 56 Ph.D., 37 Ph.D.
students)
z Grants Ph.D and Dr.Sc. (Technical) degrees

URL: http://www.spiiras.nw.ru

6
SPIIRAS

SPIIRAS Research Directions

z Fundamentals of the Informatization of the Society and

Regions, Regional Information and Computer Networks and
Systems
z Architecture, System Decisions and Software Development for
Information and Control Complexes for Real Time Signal
Processing
z Fundamentals of Information Processes in Complex (Socio-,
Eco-, Bio-,Geo-, etc) Systems
z Theoretic Basics in Developing Information Technologies for
Research Automation, Control and Manufacturing Intelligent
Systems

7
SPIIRAS STRUCTURE
SPIIRAS
Dissertation
Scientific Council for Doctor
Council
DIRECTOR
Yusupov Rafael M.,Dr..Sci., Prof. of Sciences
Degree

Deputy-Directors Deputy-Directors Deputy-Directors for

Deputy-Director for Scientific Secretary Assistant for International
for Research for Research Innovation
Maintenance Backuradze Dmitry Research Cooperation
Smirnov Alexandr V. Sokolov Boris V. Popovich Vasily V.
Tkach Anatoly F.,PhD. V., PhD. Podnozova Irina P., MS
Dr.Sci.Prof. Dr.Sci.Prof. Dr.Sci.Prof.

Service Post-Graduate Scientific Information Service Quality Management Museum of Carl May
Departments Courses Library Group Bureau school

Distributed Computational Computer Aided Integrated Applied Informatics

Structures Systems
Torgashev Valery A.Dr.Sci.,Prof. Smirnov Alexandr V.,Dr.Sci.Prof. Yusupov Rafael M.,Dr.Sci.,Prof. L
A
Data Transfer Systems and Information Technologies for
Computer Networks Intelligent Systems Gorodetski B
Control and Robotics
Losev Gennady M. PhD. Vlalimir I., Dr.Sci.,Prof. Timofeyev Adil V. V.Dr.Sci.,Prof. O
Object-Oriented Geo-Information
R
Information-Analytic Technologies
Systems Modeling Automation for Economics A
Popovich Vasily V., Dr.Sci. Marley Vladimir E. Dr.Sci. Lysenko Igor V., Dr.Sci.,Prof. T
Computer and Information O
Research Automation Information Technologies for
Systems and Problem of
Ecology and Physics Research R
Information Protection Alexandrov Viktor V. Dr.Sci.,Prof.
Vorobyov Vladimir I. Dr.Sci.,Prof. Nesterov Mikhail.M.Dr.Sci.,Prof. I
E
Software Engineering and
Software Systems
Biomedical Informatics S
Skirtil Viacheslav I., PhD. Roudnitsky Sergey B., Dr.Sci.,Prof.

Information Technologies for

Information Security Clinical Biophysics
Problems of Society Informatization Pavlovski Vladimir F. , PhD.
Kotenko Igor V. Dr.Sci.,Prof.
Zabolotsky Vadim P., Dr.Sci.
G
R
Information Control Systems Information Technologies for Systems
Ponomarev Valentin .M. Dr.Sci.,Prof.
O
Analysis
Sokolov Boris V. Dr.Sci.Prof. U
Educational Information Technologies
Speech Informatics P
Grigorijva Alla I., PhD.
Ronzhin Andre L., PhD. S
CAIS Lab Research Collaboration History SPIIRAS

with US DoD Organizations

z Ontology-Driven Information Integration from Heterogeneous
Sources for Operational Decision Making Support (US Office of
Naval Research and US Air Force Research Laboratory, 2005-2006
– CRDF’ project RUM2-1554-ST-05): Case Study – Humanitarian
Logistics.
Due to this project SPIIRAS is the first (and currently the only one) Russian organization
involved into joint research of US ONR and AFRL

z Mathematical Basic of Knowledge Discovery and Autonomous

Intelligent Architectures: Knowledge Fusion in the Scalable
Infosphere (US AFRL, 2000-2003 – ISTC’ project 1993P): Case
Study – Mobile Hospital Configuration

9
SPIIRAS

Collaboration with Ford

z Ontology Modeling and Knowledge Integration for Supply Chain

Management and Product Lifecycle Management (Ford Research Lab,
Dearborn, USA, 2001-2005)
z External Logistics Network Configuring for Russian Assembly Plant
(Ford Motor Company – Russia, St.Petersburg, Russia, 2001-2002)

z Customer-Oriented Management of Vehicles Supply Chain Using Fuzzy

Coalition Games (Ford Research Center, Aachen, Germany, 1999–
2000 - project)
z Configuration and Optimization of Global Production Networks in Order
to Improve Investment Efficiency over Total Facility Life-Time (Ford
Research Center, Aachen, Germany, for 1996-1999 - project)

10
SPIIRAS

CAIS Lab’ Current European Grants & Projects

z ILIPT - Intelligent Logistics for Innovative Product Technologies (European

Community – Research Program on Information Society Technologies,
2004-2007, – project IST-2002-507592).
Due to this project SPIIRAS is the first (and currently the only one) Russian organization involved into EU 6th FP
projects related to the business area
z IMS-NoE – Intelligent Manufacturing Systems (European Community –
Research Program on Information Society Technologies, 2003-2006, –
project IST-2001-65001). Due to this project SPIIRAS is the first (and currently the only one) Russian
organization involved into Intelligent Manufacturing Systems Program
z Knowledge Supply for Regional and Inter-Regional Networks of Small and
Medium-Size Enterprises (Swedish Foundation for International Cooperation
in Research and Higher Education, 2003-2006)
z Information Modelling for Multi-Lingual System Development Across the
Extended Enterprise and Multi-Agent Systems (Engineering and Physical
Sciences Research Council, UK, 2003-2005)
z Ontology-Based New Order Code Generation for Corporate Product Data
Management System (Festo, Germany, 2005-2006)
z Ontology-Based Intelligent Access to Documents and Catalogues (Festo,
Austria-Germany, 2003-2005)

11
SPIIRAS

CAIS Lab’ Current Russian Grants & Projects

z Methodological and Mathematical Foundations of Context-Driven
Intelligent Decision Support Systems Development (Russian Basic
Research Foundation, 2005-2007 - project 05-01-00151)

z Context-Driven Methodology of Distributed Systems Development for

Intelligent Decision Making Support in Open Information Environment
(Presidium of Russian Academy of Sciences – Research Program on
Mathematical Modeling and Intelligent System, 2004-2006 - project
16.2.35)

z Theoretical Foundations and Multi-Agent Technology for Context

Management in Open Information Environment (Department of Information
Technologies and Computational Systems of Russian Academy of
Sciences – Research Program on Fundamental Basis of Information
Technologies and Systems, 2003-2005, - project 1.9)

12
SPIIRAS

Selected Publications

z Smirnov A., Pashkin M., Chilov N., Levashova T. and A. Krizhanovsky

Agent-Based Intelligent Support to Coalition Operations: a Case Study
of Health Service Logistics Support. Information & Security. Special
Issue on IT in Coalition and Emergency Operations. Vol.16, 2005, pp.
41-61.
z Smirnov A.V., Pashkin M.P., Chilov N.G., and Levashova T.V.
Knowledge Logistics in Information Grid Environment. Future
Generation Computer Systems, 2004, 20 (1), pp. 61—79.
z Smirnov A., Pashkin M., Chilov N., Levashova T. KSNet-Approach to
Knowledge Fusion from Distributed Sources. Computing and
Informatics. V. 22, 2003, pp. 105—142.
z Smirnov A.V., Pashkin M.P., Chilov N.G., Levashova T.V. Agent-Based
Support of Mass Customization for Corporate Knowledge
Management. Engineering Applications of Artificial Intelligence,16 (4),
June 2003, pp. 349—364.
z Smirnov A., Pashkin M., Chilov N., Levashova T. Haritatos F.
Knowledge Source Network Configuration Approach to Knowledge
Logistics. International Journal of General Systems, 2003, 32 (3), pp.
251—269.

13
Introduction: SPIIRAS

Network-Centric Operations

z Network-centric Operations exploit information and network

technologies to integrate widely dispersed human decision-
makers, networking sensors, and resources into a highly
adaptive, comprehensive network-centric environment to
achieve shared situation awareness and unprecedented mission
effectiveness by efficient linking knowledgeable components in
the environment

(Adapted from the Chief of Naval Research’ definition of NCW, ONR BAA 05-013)

14
Introduction: SPIIRAS

Application Domains

z Emergency preparedness and response (to terrorism attacks /

incidents, catastrophic events, natural disasters, emergency
situations, etc.)
z Global war on terrorism (GWOT) and Multinational operations
other than war (OOTW)
z Intelligent transportation systems
z Supply chain management & e-Business
z Coalition health service logistics support
z …

15
Introduction: SPIIRAS

Decision Level Correspondence

z Strategic decisions concern general directions, long-term goals

and relationships.
z Tactical decisions take place within the context of strategic
decisions. They are primary concerned with the most appropriate
effective use of available resources.
z Operational decisions affect activities taking place right now. The
tasks, resources, and goals of these activities have been set by
strategic and / or tactical decisions.

16
Introduction: SPIIRAS

Operational Decision Making Features

z Coordination of different levels of decision making

z Management of huge amount of information and knowledge
z Intelligent sharing and reuse of information and knowledge
z Dynamic conditions
z Personalization of decision making (Intelligent Personalized Assistant)

z Decisions:
z Problem (situation) specific
z Timely
z Alternative
z Repeatable

17
Introduction: SPIIRAS

Operational DSS Reguirements

z Robustness: the decision support system (DSS) should continue to

operate even if some of its elements stop;
z Sensitivity and Adaptability: relationships between the system's
elements organization's units has to be able to be easily and quickly
readjusted in accordance with changes in the environment;
z Intensive Knowledge / Information Exchange between the system's
elements: knowledge management de facto has become essential for
decision making processes.

18
Introduction: SPIIRAS

Five Key Characteristics of Team

z Teams exist to achieve a shared goal.

z Team members are interdependent regarding some common
goal.
z Teams are bounded and stable over time.
z Team members have the authority to manage their own work and
internal processes.
z Teams operate in a social system context.

Sources:
z Alderfer C. Group and Intergroup Relations. In: J. Hackman and J. J. Suttle (eds.) Improving
Life at Work. Palisades, CA: Goodyear, 1977
z Hackman J. Introduction: Work Team in Organizations: An Oriented Framework. In: J.
Hackman (ed.) Groups That Work and those That Don’t. San Francisco, CA: Jossey-Bass,
1990

19
Introduction: What Kind of Networks Is SPIIRAS

Needed for Operational Decision Making?

z Social networks
z who knows whom => Virtual Communities
z Knowledge networks
z who knows what =>Human & Knowledge Management
z Information networks
z who informs what => Internet/Intranet/Extranet/Grid
z Work networks
z who works where => GroupWare
z Competency networks
z what is where => Knowledge Map
z Inter-organizational network
z organizational linkages => Semantic-Driven Interoperability

20
Introduction: Importance of Semantic-Driven SPIIRAS

Interoperability

z Today the organizational competence is the function of collective

intellectual capital (knowledge) of the network-centric environment
across the decision lifecycle
z Decision knowledge is critical core competency for future. Only 20% of
a firm’s knowledge is effectively used by today’s organizations.
z Different consumers (decision makers) of decision information look at it
from different contexts (aspects)
z The major problem of modern DSSs is to provide a unified and
complete view of all aspects of decision making to provide team
decision

21
Introduction: SPIIRAS

Network-Centric Environment Basis

ΠIntelligent Agents,
ΠOntology Management, and
ΠMarkup Languages service
ΠOpen Service Reference Model:
service owner 1 consumer
marketplace 2
market owner
service owner 2 service contract

marketplace 1
marketplace 3

service owner 3
22
Knowledge Logistics:
SPIIRAS

Definition

z Aim
z Acquisition, integration, and transfer of the right knowledge from
right sources in the right context to the right person in the right
time for the right purpose (6R or 6Right)

z Conditions
z Individual user requirements (personalisation),
z Available knowledge sources (information fusion),
z Current situation analysis (context)

23
Knowledge Logistics: SPIIRAS

KSNet-Approach
z Network of knowledge sources located in information environment is
referred to as “Knowledge Source Network” (KSNet)
z KSNet originates from the concept “Virtual Organization” based on the
synergistic use of knowledge from multiple sources

G H Knowledge Source Network

B
F K
(KSNet)
A
D
I
E
C
• Knowledge Bases
• Experts
• Repositories Elements of the Network
• Tools

G H G H
B
F K F K
A D
I I
X E E tn
C
1 Uni
t 2
... Uni
Outsourcing Unit
Competitors
G HG
B B B
K F F K
A A DA D
D I I I
E C C E
Partner 1 Partner 2 Partner n

24
Knowledge Logistics: SPIIRAS

FIPA Ontology Definition

z Ontology is an explicit specification of a structure of a certain domain
z Ontology includes a vocabulary for referring to a subject area, and a set of
logical statements expressing the constraints existing in the domain and
restricting the interpretation of the vocabulary
z Ontology provides a vocabulary for representing and communicating knowledge
about some topic, and a set of relationships and properties that hold for the
entities denoted by that vocabulary

z “I would say that all practical ontologies are semiformal, and the "sweet spot" is
an ontology that specifies clearly how you can commit to it. Both the formal and
informal parts should be designed to make it easy to play by the rules: the
formal by automated testing and the informal by well-written documentation” –
T. Gruber

(Thomas Gruber' interview available in AIS SIGSEMIS 1(3) 2004

http://www.sigsemis.org/newsletter/october2004/tom_gruber_interview_sigsemi)

Source: Foundation for Intelligent Physical Agents (FIPA), www.fipa.org

25
Knowledge Logistics: SPIIRAS

Ontology-Driven Methodology
Knowledge
Map
User – Logistics Manager
Knowledge Source Ontology
Parametric
Constituent Application Knowledge
User Request User
Request Ontology Ontology Source
Structural Request Constituent Constituent
Profile Constituent Application Processing
User
Ontology
Constituent
Constituent Correspondence
Answer

Correspondence Knowledge
Knowledge SourceSources
Instances
User – Operational Application
User –
Manager Ontology
Engineer

Request
Request Passive Active
Ontology
Ontology Sources Sources
Profile
Domain
Domain Tasks
Tasks&& Profile
Ontology
Ontology Methods
Methods
Ontology
Ontology

KL – Knowledge
Logistics

26
Knowledge Logistics: SPIIRAS

Ontology-Driven Knowledge Sharing

Object-oriented constraint networks as
topic-independent fundamental model
Knowledge Model primitives Universal
representation abstraction
level Knowledge representation language level
(e.g., KIF, DAML+OIL, OWL)
Language primitives

Ontology library Shared

Domain ontologies Tasks & methods abstraction
Knowledge ontologies level
sharing
level Application Specific
ontologies abstraction
level

Knowledge map
Knowledge Individual
ownership Knowledge sources including humans abstraction
level level

27
Context-Driven Methodology: SPIIRAS

Context Definition

z Context is any information that can be used to characterize

the situation of a component, where a component can be a
person, place, physical or computational object.

z For problem solving “context is what constraints a problem

solving without intervening in it explicitly” (Brézillon 1999).

Resource:
Brézillon P., “Context in problem solving: A survey”, The Knowledge Engineering Review, vol. 14, no. 1,
1999, p. 1—34.

28
Context-Driven Methodology: SPIIRAS

Motivation

Decision
Quality

Information
Volume

No All
Right Context

Theorem 1: 50%of the problems in the world result from people using
the same words with different meanings.
Theorem 2: the other 50% of the problems results from people using
different words with the same meaning.

Source: Kaplan S. The Words of Risk Analysis, Risk Analysis, Vol.17, N 4, August 1997

29
Context-Driven Methodology: SPIIRAS

Core Message

z Contextual interpretation & integration of available mission-

focused information for operational decision making is a key
point to achieve effectiveness of network-centric operation
mission based on the Knowledge Logistics Tenet:

“The right information from right sources in the right context

to the right person in the right time for the right purpose
(operational situation)”

30
Context-Driven Methodology: SPIIRAS

Levels of Integration of Knowledge and Information

z Domain level
z Integration of heterogeneous knowledge describing the domain
knowledge
z Task level
z Integration and formalization of tasks and problem-solving
methods
z Context level
z Integration of information and knowledge relevant to the
problem or situation

31
Context-Driven Methodology: SPIIRAS

Context Types

z Abstract Context z Operational Context

information and knowledge instantiation of abstract context
relevant to the request provided by values provided by context
by context components components included in the
abstract context

Request

One-to-one
Context Component
components representation

Information and knowledge

Values relevant to the request

Operational context Abstract context

32
Context-Driven Methodology: SPIIRAS

Relation to External Sources

External sources Domain ontology

External source1 User
Property1
Propertyn
…
External sourcek Environment
Property1
External
Propertym source’s data

Data Information Knowledge

33
Context-Driven Methodology: SPIIRAS

Simon’s Model

z “Intelligence phase” is defined as “searching the environment for

conditions calling the decision” (Simon 65). The phase represents the
start of the decision process. It involves the recognition of a problem
which requires a decision, and gathering and an analysis of information
concerning the problem.
z “Design phase” is described as “inventing, developing and analysing
possible courses of actions” (Simon 65). This phase entails generation
of alternative ways aimed at a goal achievement.
z In the “choice phase” all alternatives are searched, evaluated and one
chosen as a recommended solution (Simon 65). This phase supports
the operational level of decisions.

Resource: Simon H.A., The Shape of Automation. New York: Harper & Row, 1965.

34
Context-Driven Methodology: SPIIRAS

Simon’s Model and Proposed Approach

Simon’s phase Intelligence Design Choice

names

Phase Problem recognition Alternatives Efficient alternatives

content generation selection

Steps • fixing goals • designing • evaluation &

• setting goals alternatives choosing alternatives

Proposed • abstract context • constraint-based generating efficient

approach steps composition alternatives
• operational context
producing

35
Context-Driven Methodology: SPIIRAS

Technological Framework
context management
ontology management constraint satisfaction context
management

Decision Relevant Current situation

maker information source information
reference values
Request
Request problem
definition Catalogue of
e-documents

Ontology Abstract Operational Problem Capture of

library context context solving decisions
Context-sensitive access
reference Knowledge-based Instantiated Set of problem
problem model problem model solutions
Information
source
Relevant knowledge
Decision
36
Context-Driven Methodology: Common SPIIRAS

Knowledge & Problem Representation Model

Object-oriented Constraint
Ontology constraint network satisfaction
(OOCN) problem

Class Object (class)

A set of variables
Attribute Variable

Attribute range Domain Domain

Axioms and relations Constraints Constraints

37
Context-Driven Methodology: SPIIRAS

Ontology Representation Model

z OOCN = (O, V, D, C) z C = C1 U K U C 6
z O – a set of objects z C1: (class, attribute, domain)
z V – a set of variables triple
z D – a set of domains z C2: hierarchical (“is-a”,

z C – a set of constraints “part-of”) relationships

z C3: classes compatibility
OOCN Ontology Model z C4: associative relationships
Set of objects Class z C5: attribute cardinality

Variable Attribute restriction

z C6: functional relations
Domain Attribute domain
(range)
Set of Set of relations
constraints
* Object-oriented constraint network 38
Context-Driven Methodology: SPIIRAS

Two-Stage Scenario

Preliminary Stage

z Creation of models for components of DSS

z Linkage of domain knowledge and information sources

Decision Making Stage

z User request recognition
z Context-based problem modelling
z Constraint-based problem solving
z Decision making by the user
z Capture of decisions

39
Context-Driven Methodology: SPIIRAS

Models for Components of DSS

Component of DSS Model

Ontology
Domain knowledge
Tasks &
Domain
methods

User (decision maker) User profile

Information source
Environment
capability models
Formalism of OOCN
40
Context-Driven Methodology: SPIIRAS

Linkage of Domain Knowledge with Environment

Information source Domain ontology

model
data from
Monitoring system information
sources
is a is a
Weather Weather Weather
station radar
Wind Air temperature
velocity K
Air Wind Air pressure
temperature velocity
associative
class attribute relationships

41
Context-Driven Methodology: Ontology-Based SPIIRAS

Integration of Information and Knowledge

User request Domain ontology Tasks & methods ontology Ontology slice
1
recognition Route Routing
problem 3
Request
Airline
Air availability Slicing
terms 2
Road
Search Land
Weather availability
properties Availability
Airport
Route
Location availability
Weather Monitoring
properties Weather system
radar

Merging
Information source Weather

5
representation station
Air
temperature Wind 4
Information
velocity Slicing
source slice
42
Context-Driven Methodology: SPIIRAS

Constraint-Based Problem Modelling

Pa = (O, A, Do, C),

Constraint-based Abstract
P – problem
problem model context
modelled by abstract context
O – set of classes
is-a
A – set of class attributes
Do – set of attribute domains
defined in ontology slice
C – set of constraints
Constraint-based
Operational
problem model with
context Po = (O, A, Di, C),
problem data
Po P – problem
modelled by operational context
Di – set of attribute domains of
Constraint Constraint size equal to data value provided
satisfaction problem solver by information sources
43
System “KSNet” Research Prototype: SPIIRAS

Integrated Framework

Operational decision making

Abstract Operational Problem

Problem level Request
context context solving

Components Ontologies and Ontology & Constraint

Technological representation information context satisfaction
level and knowledge sources linkage management problem
accumulating
Implementation
level Web services, Agents

44
System “KSNet” Research Prototype: SPIIRAS

Adaptive Service

User request
defines defines

representation
Abstract context Domain and
Information & task
knowledge ontologies
sources Operational context
data values

Code generation
- Prepared in advance templates
- Generated "on-the-fly" code fragments

Compilation

Execution
(Constraint Solver)

Result
45
System “KSNet” Research Prototype: SPIIRAS

Agent-Based Architecture

Service Ontology hosts distributed in the Web

Requestors Library connections between hosts
via TCP/IP
OKBC – Open Knowledge
Open Web Ontology Management Base Connectivity
Service Agent
Constraint
solver
User Context
Users Agents Management
Agent
Other OKBC
Agents Wrapper
Server
Monitoring
Expert
Agent
Assistant Wrappers External
Agents Knowledge
Facilitator, Configurator, Representation
Translation agent Systems
Experts External
Sources
46
System “KSNet” Research Prototype: SPIIRAS

Agents Negotiation Protocol Choice Criteria

z Contribution: agents have to make the best contribution

into the overall system's benefit – not into the agents'
benefits
z Task performance: the main goal is to complete the task
– not to get profit out of it
z Mediating: in all negotiation processes there is an agent
managing the process and making a final decision
z Trust: the agents can completely trust each other
z Common terms: the agents use common terms for
communication

47
System “KSNet” Research Prototype: SPIIRAS

Comparison of Negotiation Protocols

Protocols Protocols:
Criteria ΠVoting (VP)
VP BP AP MMP CGP CNP ΠBargaining (BP)
Task ΠAuctions (AP)
performance ; ; … … …/; ;
ΠCoalition Games
Contribution …/; ; … … … …/; (CGP)
ΠGeneral
Mediating … … ; … … ; Equilibrium Market
Mechanisms
Trust ; ; ; ; … ; (MMP)
ΠContract Nets
Common terms ; ; ; ; ; ; (CNP)

; – supported
…/; – weakly supported
… – not supported

48
System “KSNet” Research Prototype: SPIIRAS

Contract Net Protocol (CNP)

Manager Contractor

z More distributed negotiation than 1. Call for Proposals

MMP
1.1. Refuse
z The main features of this protocol
are: 1.2. Not Understood
1. managers (initiators in FIPA)
divide tasks, 1.3. Propose
2. contractors (participants in FIPA)
bid, 1.3.1. Reject Proposal
3. manager makes contract for
lowest bid, 1.3.2. Accept Proposal

4. there is no negotiation of bids. 1.3.2.1. Inform

1.3.2.2. Failure

UML sequence diagram of

FIPA-based contract net
protocol

49
System “KSNet” Research Prototype: SPIIRAS

Major Idea of Constraint-Based CNP

z A generic call for proposals from a manager to contractors:

Objective (optional) E.g. time→min
Constraints (optional) E.g. costs≤$20
Content (required)
z A generic contractors’ proposal:
Constraints (optional) E.g. costs=$15
Content (required)
If contractors cannot meet the requirements of the manager
they propose the closest possible parameters and manager
decides whether to accept the proposal or not.

50
System “KSNet” Research Prototype: SPIIRAS

Constraint-Based CNP Schema

Configuration Agent Wrapper
(Manager) (Contractor)
Call for Proposals

Refuse
Propose Here and then:
CSP-Based Reject Proposal Yellow ( )–
conventional
Analysis Accept Proposal CNP
Inform Green ( )–
modification
Failure

Given Ti=1…n – response time of contractors (n – is the number of

participating contractors), Tman – manager’s response time, and TC,
TM – negotiation time for conventional and modified constraint-based
CNP respectively, TC ≤ TM ≤ TC + Tmax + Tman holds, where Tmax = Ti.
51
System “KSNet” Research Prototype: SPIIRAS

Experimentation with Constraint-Based CNP

Conventional CNP Sequential CNP Constraint-Based CNP

time → min time → min time → min

time → min 30 min/$15 time → min
time → min costs → min AND time ≤ 30 time → min
15 min/$20 25 min/$10 15 min/$20
30 min/$15 time ≤ 30 AND costs → min 30 min/$15
50 min/$25 50 min/$25 50 min/$25
time ≤ 50 AND costs → min
time ≤ 50 AND costs → min
30 min/$15
45 min/$5

50 min and $60 50 min and $50 50 min and $45

52
System “KSNet” Research Prototype: SPIIRAS

Comparison with Other Multiagent Systems (MAS)

Project KRAFT (UK) OBSERVER (SPAIN) InfoSleuth (USA) KSNet (Russia)

MAS for information search System for synonymic Open MAS for information search MAS for rapid Ontology-Driven
under user defined constraints relations based information Information Integration from
Project Goal search distributed heterogeneous
sources for Decision Making

KQML, P/FDM, CoLan, CIF Internal formats OKBC, LISP, CLISP, LDL+ OWL, DAML+OIL, KQML
Formats and OOCN
Standards

N/A Not supported Requests history User profiles, request

Information
ontologies
about Users

Base Ontologies WordNet Application (domain specific) Application (domain specific) Application (domain specific)
ontologies ontologies ontologies

Hierarchy Lattice Meta-level ontology, Top-level ontology, taxonomy,

Ontologies hierarchy
Implicit hierarchy
organization

FIPA-based with peer-to-peer Not supported FIPA-based with mediating FIPA-based with peer-to-peer &
MAS
interaction interaction mediating interaction
Architecture

53
System “KSNet” Research Prototype: SPIIRAS

Standards of Information Kernel

Ontology
Management

User DAML+OIL
Interface
RDF Representation
Data HTML XML
Management
ISAPI / CGI
ODBC Implementation
Internal Object Scheme
Classes Attributes Constraints Domains
Notation
Object-Oriented Constraint Networks

DAML – DARPA Agent Markup Language HTML – HyperText Markup Language

OIL – Ontology Inference Layer ISAPI – Internet Server Application Programming Interface
RDF – Resource Description Framework CGI – Common Gateway Interface
XML – Extensible Markup Language ODBC – Open DataBase Connection

54
System “KSNet” Research Prototype: SPIIRAS

Standards of Service-Oriented Model

Semantics
Description and discovery UDDI Ontology
Service connection WSDL
Message protocol SOAP
Message syntax XML
HTTP
Transport
TCP/IP
UDDI - Universal Description, Discovery, & Integration – a “meta service” for locating open services by
enabling robust queries against rich metadata;
WSDL - Web Services Description Language – Interface Definition Language for open services;
SOAP - Simple Object Access Protocol – XML-based RPC protocol;
XML - Extensible Markup Language – a specification developed by the W3C for a pared-down version of
SGML;
HTTP - HyperText Transfer Protocol – the underlying protocol used by the Internet;
TCP/IP - Transmission Control Protocol/Internet Protocol – the suite of communications protocols used to
connect hosts on the Internet.

55
Case Study: SPIIRAS

Humanitarian Logistics

z The number of annual natural and human-made disasters has tripled

since 1970. The strains on humanitarian organizations responding to
emergencies showed that in last year 256 million people were reported
affected by disasters, while the annual average is 210 million. The
practice shows that one of the most difficult steps is getting the right
relief supplies to the people in need at the right time. At the same time
delivering of too much supplies or wrong supplies means loosing time
and money. Therefore, humanitarian logistics standing for processes
and systems involved in mobilizing people, resources, skills and
knowledge to help vulnerable people affected by natural disasters and
complex emergencies, is central for disaster relief

Source: Humanitarian Logistics: Getting the Right Relief to the Right People at the Right
Time, Fact Sheets, Fritz Institute, 2005 URL: http://www.fritzinstitute.org/-fact_sheets/-f_s-
hls.html.

56
Case Study: SPIIRAS

Requested Information

z Hospital related information z Geography and weather of the

(structure, components, times of Binni region (types, routes, and
delivery) time of delivery in dependence
z Available United Nations and on a method of delivery, e.g. by
friendly suppliers (suppliers’ air, by trucks, by off-road
capabilities, capacities, vehicles)
locations) z Political situation, e.g. who
z Available United Nations and occupies used for transportation
friendly providers of territory, existence of military
transportation services actions on the routes, etc.
(available types, routes, and (additional constraints on routes
time of delivery) of delivery)

57
Case Study: SPIIRAS

Ontology Creation

Internet Knowledge Sources

Experts Experts
Request vocabulary List of Knowledge Sources
Hospital, structure, - Catalogue of hospital equipment:
components, times of delivery, http://maktechno.virtualave.net/m
suppliers, capabilities, enu.htm
capacities, locations, providers, - Naics.daml (DAML Ontology
Request transportation services, types, Library)
routes, geography, weather, - CLIN-ACT (Clinical Activity).html
KSNet method of delivery, delivery (The ON9.3 Library of
Ontologies)
- Medical Procedures.daml (The
ON9.3 Library of Ontologies)
- Upper Cyc/HPKB IKB
Ontology.html (Ontolingua
Ontology Creation Server)

58
Case Study: SPIIRAS

Complex of Tasks

z Treatment Course Definition

z find the right treatment course for the given injury type
z Hospital allocation
z find the most appropriate location for a portable hospital considering
locations of the disaster, water resources, nearby cities and towns,
communication facilities (e.g., locations of airports, roads, etc.), etc.
z Resource allocation
z find the most efficient hospital configuration considering type and quantity
of material or goods required for the hospital, properties (location, costs,
productivity, availability, etc.) of suppliers of the materials and goods,
optimization parameters (costs or time)
z Routing problem
z find the most efficient ways of delivery of the hospital's components from
available suppliers considering communications facilities (e.g., locations of
airports, roads, etc.), their conditions (e.g., good, damaged or destroyed
roads), weather conditions (e.g., rains, storms, etc.)

59
Case Study: SPIIRAS

Tasks & Methods Ontology

Disaster relief
• type of disaster
part of

Hospital allocation Routing problem Resource Treatment course

• position • current situation allocation definition
• city, disaster type, conditions • supplier • type of disaster
infrastructure • list of suppliers properties ¾ injury types
facilities • location ¾ list of ¾ type and quantity
• possible location ¾route availability suppliers of medical facilities
¾location

part of part of
Notation:
Hospital location Supplier definition
• current situation Task • type and quantity of
conditions • input argument medical facilities
¾possible location ¾output argument ¾supplier properties
60
Case Study: Interactions between Domain and SPIIRAS

Tasks & Methods Ontologies

Domain ontology Tasks & methods ontology
is a
Thing Disaster relief
is a • type of disaster
is a
part of
Transportation
Supplier o destination
Routing problem
o location
o location • current situation
o route type
o goods conditions
o dimensions
o quantity • list of suppliers
o weight
o time • destination
o time
o costs ¾route availability
o costs
part of
associative
relationship Weather factors
Weather calculation
o air pressure
o air temperature weather properties
Notation: • route type
o attribute of domain o wind velocity
oK ¾weather type
ontology class ¾coefficient

61
Case Study: Domain Ontology Slice for SPIIRAS

“Hospital Allocation” Task

Thing

Structures, building Communication

Construction and construction com- facilities
ponents and supplies

Portable
hospital

Drugs and Professional and

Staff pharmaceutical commercial equipment
products … and supplies

Suppliers

is-a part-of associative

62
Case Study: SPIIRAS

Common Application Ontology

Common Application Ontology

Treatment Course Hospital

Definition Allocation

Hospital Hospital
Components Location

Suppliers Locations
Routing

Resource Allocation

Hospital Allocation – Task; Hospital Location - Ontology Element

63
Case Study: SPIIRAS

Application Ontology Fragment

Thing
Legend:

Hospital Portable Disaster Transportation Domain

configuration Ontology Class
hospital
Task & Method
Medical Furniture Structures, building Ontology Class
equipment and construction
components is-a relationship

Suppliers part-of
relationship
associative
relationship
BOM – Bill of Materials
Components Hospital Logistics
definition allocation

BOM Resource Routing

definition allocation problem

Route
availability
64
Case Study: SPIIRAS

Abstract Context Example

Hospital Component Supplier

costs type name Hospital
costs
table_req
bed_req
Component
type
price
quantity
Supplier
name
type
capacity

table_req price type Resource Delivery

start_time
end_time
Location
name
latitude
Weather
time
conditions

Allocation
quantity longitude
availability

bed_req quantity capacity Transportation

route
cost
time
Facility
type

Hospital Component Supplier

costs type name

table_req price type
bed_req quantity capacity

Delivery Location Weather Delivery

start_time
Location

name
Weather

time

Hospital
end_time latitude conditions
quantity longitude
availability

start_time name time Transportation Facility

Allocation
route
type
cost
time

end_time latitude conditions

quantity longitude
availability
Hospital Component Supplier
costs type name
table_req price type
bed_req quantity capacity

Routing
Delivery Location Weather
start_time name time
end_time latitude conditions
quantity longitude
availability

Transportation Facility Transportation

route
cost
time
Facility
type

route type
cost
time

65
Case Study: SPIIRAS

Operational Context Example

Hospital Component Supplier Location Weather

costs ? (→ min) type bed name "mak name Laki time 17.10.04
table_req 10 price 10 technologies" latitude 16.9N conditions cloudy
bed_req 14 quantity ? type bed longitude 36.8E
capacity 10 availability ?

Location Component Supplier Location Weather

name Aida type bed name "Kerry name Dado time 17.10.04
latitude 17.7N price 16 Ultrasonics Ltd" latitude 20.3N conditions sunny
longitude 35.3E quantity ? type bed longitude 37.3E
availability ? capacity 15 availability ?

Weather Component
time 17.10.04 type table
conditions sunny price 9 …
quantity ?

…
66
Case Study: SPIIRAS

Examples of Constraints

Œthe attribute costs (q1) belongs to the class hospital (o1):

cI1 = (o1, q1);
Œthe attribute costs (q1) belonging to the class hospital (o1) takes
positive values: cII1 = (o1, q1, R+);
Œ mobile hospital (o5) is a hospital (o1): cIV1 = 〈o1, o5, 0〉;
Πthe class medical equipment (o4) is a part of the class hospital (o1):
cIV1 = 〈o1, o4, 1〉;
Πthe class furniture (o2) is compatible with the class furniture supplier
(o3): cIII1 = ({o2, o3}, True);
Πthe attribute capacity of the class mobile hospital (o5) serves as an
input parameter in the class components definition (o6): cV1 = (o5, o6);
Œthe value of the attribute cost (q1) of an instance of the class furniture
(o2) depends on the values of the attribute price (q2) of instances of the
class suppliers (o6) and on the number of such instances (the attribute
quantity (q3) of the class furniture): cVI1 = f({o2, q1}, {(o6, q2),
(o2, q3)});

67
Case Study: SPIIRAS

Hospital Configuration Problem

Major Features

0 Capacity
Time
Costs
Capacity
Time
Costs
Time
Costs
Hosp. Allocation
1
BOM Definition
Geographical
Injury specifics
Capacity objects

Hosp. Suppliers
Required Definition Hospital
2 Supplies Suppliers and their
capacities
Location

Used Routing

3 Suppliers Hosp. location

Suppliers locations
Path network

68
Case Study: SPIIRAS

Example Result of User Request Processing

69
Case Study: SPIIRAS

Uncertainty Treating Example

z Number of operating tables
per patient:
z Experts' estimations (min;
max): [2; 8], [4; 6], [1; 7],
[3; 6], [4; 8], [2; 7], [3; 7]
z Influence:
z If only 3 tables are
considered the decision's
reliability is 66.67%

70
Case Study: SPIIRAS

Example Result of User Request Processing

Constraint is omitted
Solution's reliability 64%
71
Case Study: SPIIRAS

Example Result of User Request Processing

Constraint is not omitted

Solution's reliability 100%

72
SPIIRAS

Conclusions (1)

z Context-driven decision making is of on-the-fly agent-based intelligent

service based on integration of ontology & context management and
constraint satisfaction technologies .
z The context-driven knowledge integration approach for operational
decision support is originally problem-independent and can be applied
to different domains by creation of a new application ontology describing
the new problems, and finding and attaching appropriate information &
knowledge sources.
z Implementation of context-driven methodology can significantly facilitate
flexibility and response speed of operational decision support systems
for network-centric operations.
z Implementation of multi-agent technology together with semantic-driven
interoperability create an opportunity for fast development of scalable
DSSs.

73
SPIIRAS

Conclusions (2)

z Context-Driven Decision Making Approach has direct relationships with

following Future Directions of ONR’ Collaborative & Knowledge
Management Program:
z Decision Making through Synthetic Experience,
z Consultative Advisors,
z Context-Sensitive Filters,
z Reconfigurable Information Fusion Capability

74
Future Work: SPIIRAS

Motivation
z Fast (time-critical) response for megadisaster events (massive
hurricanes, earthquakes, nuclear attacks, etc.) requires emergency
preparedness based on long-term response scenarios planning
with realistic (or predictable) expectations concerning available
(alternative) federal & local sources and estimation of access time
to them.
z Major Megadisaster Response Management Issues:
z scenario-based information fusion for operation preparing
related to typified situation;
z context-aware interoperability of operation participants based on
common knowledge & problem representation model;
z on-the-fly decision support assistance for officials based on
adaptive services.

75
Future Work: SPIIRAS

Noncombatant Evacuation Operation

z Purpose
z Producing efficient plans for treatment and evacuation of injured people
based on information available in different sources
z Information used
z Patient data: location, injury, time of injury, current condition, personal data
(diseases, disease history, regular drug prescriptions, etc.)
z Providers of evacuation facilities: the facilities
z Weather conditions
z etc.

z Possible decisions
z treat the patient at site
z send mobile ambulance to the site

z send a helicopter to transport the patient to hospital

z etc.

76
Future Work: SPIIRAS

Extended Information Environment

Medical Data and

Knowledge Bases Geo
(e.g. BUMED bases) Information
-ID System
-detailed -information for
personal -weather conditions forecast
Microchip with diagnostics
information -geography
personal info -Identity (id) -possible courses of
(e.g. RFID) -initial personal data, (diseases, treatment and drug
medicines, prescriptions Solution
such as age, diseases
etc.)
(e.g., diabetes, allergy,
-treat at site
etc.), etc.
-use mobile ambulatory
Patient
-help at site, then
transport to naval hospital
Knowledge Fusion -transport to naval
Based Intelligent hospital immediately
-location Support System -…
-injury Schedule
-time of injury -capacities probability of
-capacities -facilities
-current condition -capacities survival
-facilities -schedule
-facilities -schedule
-schedule probability
Mobile Naval Evacuation threshold
Ambulatory Hospital Facilities
feasible time time
span

77
Future Work: SPIIRAS

Scenario

Start

Patient, Treatment Course Medical Knowledge

RFID Definition Bases

Hospital
Allocation

Medical Resource Geo Information

Facilities Allocation System

Routing

End
78
Future Work: SPIIRAS

Input of Preliminary Stage

„ Available scenarios and decision making methods / models /

pattern / rules,
„ Available information sources for decision makers,
„ “National Context” descriptions - cultural specifics (behavior,
values, etc) for different nationals, geographical regions, etc.
„ Formats of external ontologies and information sources,
„ Case study documentation

79
SPIIRAS

Thank you!

Prof. Alexander SMIRNOV (e-mail: [email protected] )

80