Designing an extended set of coordination mechanisms for multi agent systems

DESIGNING AN EXTENDED SET OF COORDINATION
MECHANISMS FOR MULTI-AGENT SYSTEMS

by
Wei Chen

A dissertation submitted to the Faculty of the University of Delaware in partial
fulfillment of the requirements for the degree of Doctor of Philosophy in Computer
Science

Fall 2005

c 2005 Wei Chen
All Rights Reserved


UMI Number: 3200551

Copyright 2005 by
Chen, Wei
All rights reserved.

UMI Microform 3200551
Copyright 2006 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.

ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, MI 48106-1346

DESIGNING AN EXTENDED SET OF COORDINATION
MECHANISMS FOR MULTI-AGENT SYSTEMS

by
Wei Chen

Approved:
B. David Saunders, Ph.D.
Chairperson of the Department of Computer and Information Sciences

Approved:
Thomas M. Apple, Ph.D.
Dean of the College of Arts and Sciences

Approved:
Conrado M. Gempesaw II, Ph.D.
Vice Provost for Academic and International Programs


I certify that I have read this dissertation and that in my opinion it meets the
academic and professional standard required by the University as a dissertation
for the degree of Doctor of Philosophy.

Signed:
Keith S. Decker, Ph.D.
Professor in charge of dissertation

I certify that I have read this dissertation and that in my opinion it meets the

academic and professional standard required by the University as a dissertation
for the degree of Doctor of Philosophy.

Signed:
Daniel L. Chester, Ph.D.
Member of dissertation committee

I certify that I have read this dissertation and that in my opinion it meets the
academic and professional standard required by the University as a dissertation
for the degree of Doctor of Philosophy.

Signed:
Chien-Chung Shen, Ph.D.
Member of dissertation committee

I certify that I have read this dissertation and that in my opinion it meets the
academic and professional standard required by the University as a dissertation
for the degree of Doctor of Philosophy.

Signed:
Matthew J. Hoffmann, Ph.D.
Member of dissertation committee


ACKNOWLEDGMENTS

I would like to thank all my colleagues in our multi-agent system group of the
Department of Computer and Information Sciences, University of Delaware, especially
the leader and my adviser, Dr. Keith S. Decker.
My deepest thanks to Keith for his guidance and support, especially for his patience and help throughout the process of my dissertation. Without these, my research

objectives would never have succeeded quickly and well. His outstanding ideas in this
research area have given rise to various projects, including this dissertation.
I would like to give very special thanks to Dr. Daniel L. Chester, who pointed out
most of the mistakes in early versions of this dissertation and provided many constructive ideas; his expertise and interest in various research fields is a great inspiration for
my life and for my future research work. Deep thanks to Dr. Chien-chung Shen for his
sharp questions and directions; I learned a lot from his character and serious attitude towards research. Many thanks to Dr. Matthew J. Huffmann for his guidance and insights
from a special perspective, which helped open up this dissertation to readers outside of
distributed artificial intelligence.
I also offer special thanks to the researchers in the MAS group of the University of
Massachusetts at Amherst. I went to the UMass campus for ideas during the early stage of
my research work and got in touch with some of their thoughts. My advisor’s advisor, Dr.
Victor Lesser, showed me the potential development in the field of multi-agent systems
and encouraged me to continue work in this research area.

iv


I should present special thanks to my wife, Guangmeng (Grace), for her continuous support and all the love, care, and help in my life. I am profoundly thankful to my
parents; I hope they will remain happy, healthy, and proud of their son.

v


TABLE OF CONTENTS

LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii
LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvi
ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii
Chapter
1 INTRODUCTION: WHAT AND WHY? . . . . . . . . . . . . . . . . . . .

1.1
1.2
1.3
1.4

What Is Coordination? . . . . . . . . . . . . . . . . . . . . . . . .
Why study coordination? . . . . . . . . . . . . . . . . . . . . . .
Coordination in Multi-Agent Systems . . . . . . . . . . . . . . . .
Coordination in a Domain Application—Emergency Medical Service
Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.4.1

1.5

2
5
8

. . . 14

Three Coordination Points in EMS . . . . . . . . . . . . . . . . 16

Intellectual Contributions . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.5.1
1.5.2
1.5.3
1.5.4
1.5.5

1.6

. . .
. . .
. . .

1

Problem Scope . . . . . . . . . . . . . . . . . . . .
Formal Representation . . . . . . . . . . . . . . . .
An Extended Set of GPGP Coordination Mechanisms
Architectural Support . . . . . . . . . . . . . . . .
Experimentation and Qualitative Analysis . . . . . .

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

.
.
.
.
.

21
23
24
27
28

Chapter Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

vi


2 RELATED WORK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.1

Coordination Approaches in Multi-Agent Systems . . . . . . . . . . . . 31
2.1.1

Coordination Science Presented From MIT . . . . . . . . . . . . 33
2.1.1.1
2.1.1.2

2.1.2

Direct Inspiring Base — TÆMS and GPGP . . . . . . . . . . . . 37
2.1.2.1
2.1.2.2
2.1.2.3

2.1.3

Ideas and Models . . . . . . . . . . . . . . . . . . . . 46
Coordination Selection Strategy . . . . . . . . . . . . . 47

Teamwork . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.1.4.1

2.1.4.2

2.2
2.3

TÆMS . . . . . . . . . . . . . . . . . . . . . . . . . 37
GPGP . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Other work related to TÆMS and GPGP . . . . . . . . 42

Jennings and Colleagues . . . . . . . . . . . . . . . . . . . . . . 45
2.1.3.1
2.1.3.2

2.1.4

Key Ideas and Coordination Methods . . . . . . . . . . 33
Applications . . . . . . . . . . . . . . . . . . . . . . . 37

Teamwork Model and Agent Architecture . . . . . . . 50
Current Work . . . . . . . . . . . . . . . . . . . . . . 51

Other Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

3 DEVELOPING AN EXTENDED SET OF GPGP COORDINATION
MECHANISMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.1
3.2

Extending Traditional GPGP Coordination Approach . . . . . . . . . . . 61

Features of Our Extended GPGP Approach . . . . . . . . . . . . . . . . 63
3.2.1
3.2.2

Task Structures and Enables Relationship . . . . . . . . . . . . . 63
Real time task structure alteration . . . . . . . . . . . . . . . . . 66

vii


3.2.3
3.3

Abstraction Level . . . . . . . . . . . . . . . . . . . . . . . . . 69

An Extended Set of GPGP Coordination Mechanisms for Enables
Relationships . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
3.3.1
3.3.2
3.3.3
3.3.4
3.3.5
3.3.6
3.3.7
3.3.8
3.3.9
3.3.10
3.3.11
3.3.12
3.3.13

3.3.14
3.3.15
3.3.16
3.3.17

Avoidable Dependency . . . . . . . . . . . . . . .
Sacrifice Avoidable Dependency . . . . . . . . . . .
Coordination by Reservation . . . . . . . . . . . .
Predecessor Earliest Start Time (EST) Commitment .
Predecessor Deadline Commitment . . . . . . . . .
Predecessor Notice at Start . . . . . . . . . . . . .
Predecessor Sending Result . . . . . . . . . . . . .
Successor Deadline Commitment . . . . . . . . . .
Successor Earliest Start Time (EST) Commitment . .
Demotion Shift . . . . . . . . . . . . . . . . . . .
Promotion Shift . . . . . . . . . . . . . . . . . . .
Polling for Result . . . . . . . . . . . . . . . . . .
Polling for Schedule . . . . . . . . . . . . . . . . .
Constant Headway / Timetabling . . . . . . . . . .
Third Party Execution . . . . . . . . . . . . . . . .
Third Party Coordinator . . . . . . . . . . . . . . .
Bidding . . . . . . . . . . . . . . . . . . . . . . .

.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

73

76
77
78
80
81
82
82
84
84
86
86
87
88
89
90
91

4 FORMAL REPRESENTATION AND AGENT ARCHITECTURAL
SUPPORT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
4.1

Extended Hierarchical Task Networks—EHTNs . . . . . . . . . . . . . . 93
4.1.1
4.1.2

4.2
4.3
4.4
4.5

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Extending HTNs (EHTN) to Represent Coordination Problems . . 96

Agent Architectural Support and A Schedule Coordination Problem . .
GPGP Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . .
Coordination Mechanism Selection . . . . . . . . . . . . . . . . . . .
The Recast of Selected GPGP Coordination Mechanisms Using EHTNs
4.5.1

.
.
.
.

103
108
112
117

Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120

viii


4.5.2

Avoidable Dependency . . . . . . . . . . . . . . . . . . . . . . 121
4.5.2.1

4.5.3

Coordination by Reservation
4.5.3.1
4.5.3.2

4.5.4
4.5.5
4.5.6
4.6
4.7

Re-writing of the Extended HTNs

. . . . . . . . . . . 123

. . . . . . . . . . . . . . . . . . . 123

Re-writing of the Extended HTNs . . . . . . . . . . . 125
Coordination Protocol . . . . . . . . . . . . . . . . . . 127

Demotion Shift Dependency . . . . . . . . . . . . . . . . . . . . 129
Coordination by Sending Result . . . . . . . . . . . . . . . . . . 130
Coordination by Polling Result . . . . . . . . . . . . . . . . . . 133

Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Summary about Formal Representation . . . . . . . . . . . . . . . . . . 135

5 EVALUATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
5.1
5.2

5.3
5.4
5.5

Modeling Task Environment . . . . . . . . . . . . . . . . . . . . . . .
General Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Adjust Coordination Behaviors Under Changing Environmental Factors
Coordination as an Independent Component . . . . . . . . . . . . . . .
Analysis of Coordination in Emergency Medical Services . . . . . . . .
5.5.1
5.5.2
5.5.3
5.5.4

.
.
.
.
.

138
139
142
142
144

.
.
.
.

144
145
147
150

Framework Overview . . . . . . . . . . . . . . . . . .
Introduction to Participating Agents . . . . . . . . . .
EMS System Operation Overview . . . . . . . . . . .
System Development . . . . . . . . . . . . . . . . . .
The Application of Coordination Mechanisms to Three
Major Coordination Points in EMS . . . . . . . . . . .
Random Distribution for Incident Generation, Agent
Working Time, and More . . . . . . . . . . . . . . . .

150
153
157
159

Background . . . . . . . . . . . . . . . . . . . .
Introduction of Emergency Medical Service (EMS)
EMS Model . . . . . . . . . . . . . . . . . . . .
Emergency Medical Service (EMS) Framework . .
5.5.4.1
5.5.4.2
5.5.4.3
5.5.4.4
5.5.4.5
5.5.4.6

ix

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.

.
.

.
.
.
.

161
163


5.5.4.7
5.5.5

System Evaluations . . . . . . . . . . . . . . . . . . . 166

Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
5.5.5.1

Response Time upon Different Number of Agents . . . 168
5.5.5.1.1
5.5.5.1.2
5.5.5.1.3

5.5.5.2
5.5.5.3
5.5.5.4
5.5.5.5
5.5.5.6

5.5.5.7
5.5.6

Three Situations: No Communication, With
Communication, and Bidding . . . . . . . . . 169
Response Time on Fixed Consecutive
Emergencies . . . . . . . . . . . . . . . . . 170
Response Time on Simulated Emergencies . . 172

Base Case for the Application of Mechanisms to
Coordination Points in EMS . . . . . . . . . . . . .
Experiment Setup . . . . . . . . . . . . . . . . . . .
Effects on Mechanisms Applied to Coordination Point
One . . . . . . . . . . . . . . . . . . . . . . . . . .
Effects on Mechanisms Applied to Coordination Point
Two . . . . . . . . . . . . . . . . . . . . . . . . . .
Effects on Mechanisms Applied to Coordination Point
Three . . . . . . . . . . . . . . . . . . . . . . . . .
Analyzing Entire Coordination Process of EMS . . .

. 174
. 177
. 178
. 182
. 183
. 184

Future Work on the EMS Framework and Experimentation . . . . 187

6 CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188

6.1

Limitations of This Work . . . . . . . . . . . . . . . . . . . . . . . . . 189
6.1.1
6.1.2

6.2

Self-Interested Agents . . . . . . . . . . . . . . . . . . . . . . . 189
Coordination Among Heterogeneous Agents . . . . . . . . . . . 189

Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
6.2.1
6.2.2
6.2.3
6.2.4
6.2.5
6.2.6

Two General Goals for Future Research . . . . . . . . . . .
Coordination in Cascading Situations . . . . . . . . . . . .
Future Work about EMS and Potential Extension to Manage
Large-Scale Emergencies—Disasters . . . . . . . . . . . .
Discussion about the Facilitates Task Relationship . . . . .
Facilitates and Hinders . . . . . . . . . . . . . . . . . . . .
Application in Limited Resource Problems (LRPs) . . . . .
x

. . . 190
. . . 191

.
.
.
.

.
.
.
.

.
.
.
.

192
194
196
196


6.2.7

Application in Bioinformatics—An Example Domain of LRPs . . 198

BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
Appendix
A PROOF OF EXPRESSIVENESS THEOREM . . . . . . . . . . . . . . . . 214
B TASK STRUCTURE EQUIVALENCE AND REDUCTION . . . . . . . . . 218
C DISCUSSION OF INCIDENT GENERATION AND SELECTION . . . . . 221

xi


LIST OF FIGURES

1.1

An example showing a practical coordination problem. . . . . . . . . .

1.2

EMS coordination point one: between a dispatcher and a police agent. . 16

1.3

EMS coordination point two: between a police agent and an ambulance. 17

1.4

EMS coordination point three: between ambulance and hospital. . . . . 18

2.1

The Grid World Scenario. . . . . . . . . . . . . . . . . . . . . . . . . 47

2.2

System model for sensible agents. . . . . . . . . . . . . . . . . . . . . 53

2.3

The logical coordination service. . . . . . . . . . . . . . . . . . . . . 56

3.1

Example Agent Task Structure. . . . . . . . . . . . . . . . . . . . . . 67

3.2

One example of the rewrite of task structure for Coordination relation
ships. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

3.3

Avoidable Dependency. . . . . . . . . . . . . . . . . . . . . . . . . . 73

4.1

Three kinds of links connecting provisions and outcomes of the tasks
n, ni and nj . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

4.2

Original DECAF agent architecture. . . . . . . . . . . . . . . . . . . . 103

4.3

Coordination Component in an example agent architecture. . . . . . . . 105

4.4

The relationships among the planner, the scheduler, and the coordination
module. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

4.5

GPGP composition. . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
xii

9


4.6

Avoidance of interdependency-related tasks—Avoidance Coordination
Mechanism. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

4.7

The domain tasks and the inserted GPGP mechanism—coordination by
reservation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124

4.8

Sequence diagram of GPGP coordination mechanism protocol:
Reservation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

4.9

Protocol of Coordination by Reservation. . . . . . . . . . . . . . . . . 128

4.10

The domain tasks and the inserted GPGP mechanism—coordination by
demotion shift dependency. . . . . . . . . . . . . . . . . . . . . . . . 129

4.11

Sequence diagram of GPGP coordination mechanism
protocol—Demotion. . . . . . . . . . . . . . . . . . . . . . . . . . . 130

4.12

Demotion mechanism. . . . . . . . . . . . . . . . . . . . . . . . . . . 131

4.13

The domain tasks and the inserted GPGP mechanism—coordination by
SendingResult. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

4.14

Sequence diagram of GPGP coordination mechanism protocol - Sending
Result. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

4.15

The domain tasks and the inserted GPGP mechanism—coordination by
Polling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134

4.16

Sequence diagram of GPGP coordination mechanism protocol—Polling. 134

5.1

An EMS system information flow chart. . . . . . . . . . . . . . . . . . 148

5.2

A snapshot of the EMS framework demonstration program. . . . . . . . 151

5.3

EMS task structures for Ambulances, Police, and Hospitals. . . . . . . 152

5.4

State change graph for ambulances shown as a NDFA. . . . . . . . . . 156

5.5

State change graph for police shown as a NDFA. . . . . . . . . . . . . 157

5.6

Diagram of EMS operation . . . . . . . . . . . . . . . . . . . . . . . 158
xiii

5.7

Response time based on changing number of agents. . . . . . . . . . . 171

5.8

Response time based on changing number of agents (mean=20s). . . . . 173

5.9

Response time of GPGP Coordination Mechanisms on EMS
Coordination Point One. . . . . . . . . . . . . . . . . . . . . . . . . . 179

5.10

Average survival rate of selected GPGP Coordination Mechanisms on
EMS Coordination Point One. . . . . . . . . . . . . . . . . . . . . . . 180

5.11

Average coordination cost of selected GPGP Coordination Mechanisms
on EMS Coordination Point One. . . . . . . . . . . . . . . . . . . . . 181

5.12

Average response time of selected GPGP Coordination Mechanisms on
EMS Coordination Point Two. . . . . . . . . . . . . . . . . . . . . . . 182

5.13

Average survival rate of selected GPGP Coordination Mechanisms on
EMS Coordination Point Two. . . . . . . . . . . . . . . . . . . . . . . 183

5.14

Average coordination cost of selected GPGP Coordination Mechanisms
on EMS Coordination Point Two. . . . . . . . . . . . . . . . . . . . . 184

5.15

Response time of GPGP Coordination Mechanisms on EMS
Coordination Point Three. . . . . . . . . . . . . . . . . . . . . . . . . 185

5.16

Average survival rate of selected GPGP Coordination Mechanisms on
EMS Coordination Point Three. . . . . . . . . . . . . . . . . . . . . . 185

5.17

Average coordination cost of selected GPGP Coordination Mechanisms
on EMS Coordination Point Three. . . . . . . . . . . . . . . . . . . . 186

6.1

EHTN-style representation for facilitation relationship. . . . . . . . . . 195

6.2

Implementing Bioinformatics System With Multi-Agent Approach.

B.1

A real structure that represents a coordination point. . . . . . . . . . . 218

B.2

A task structure whose only purpose is to transmit messages; therefore, it
is not a coordination point. . . . . . . . . . . . . . . . . . . . . . . . 219

B.3

Equivalent task structures while different in the planning time. . . . . . 220
xiv

. . 199


B.4

Reduction of certain task structures into the coordination understandable
form. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220

xv


LIST OF TABLES

5.1

Application of extended GPGP mechanisms to the three coordination
points in emergency medical service (EMS) framework. . . . . . . . . 163

5.2

Application of extended GPGP mechanisms to the three coordination
points in emergency medical service (EMS) framework. . . . . . . . . 175

5.3

System performances caused by a combination of the best mechanisms
applied to the corresponding three coordination points . . . . . . . . . . 186

xvi


ABSTRACT

Well coordinated behaviors greatly improve intelligent entities’ overall performance,
while inappropriate coordination results in reduced system efficiency, unfinished core
tasks, misuse of key resources, and even system crashes. Thus, coordination is a central
research task in Distributed Artificial Intelligence (DAI).
This dissertation discusses our way of approaching the coordination problem,
which is to develop an extended set of coordination mechanisms to manage the interdependencies between the activities among multiple agents. This dissertation presents
four steps to tackle this problem.
First, formal representation: articulating a domain-independent approach for specifying a set of coordination mechanisms and representing the characteristics of the agents’
tasks and actions with a highly expressive formalism. We present an Extended Hierarchical Task Network (EHTN) that is expressive enough to represent worth-oriented goals,
contingencies, and the uncertainties that arise when task plans are in fact distributed over
multiple agents by annotating tasks and actions quantitatively. We design and implement an extended set of GPGP (Generalized Partial Global Planning) coordination mechanisms, which are recast using this formalism.

Second, mechanism development: constructing a large number of mechanisms to
deal with the dependencies between multiple agents’ tasks. We have catalogued seventeen GPGP coordination mechanisms for the enable relationship and discussed potential
mechanisms for the other kinds of relationships, such as facilitate and hinder.
Third, architectural support: updating the agents’ internal structure by inserting
a novel coordination module between the planner and the scheduler. We strengthen our
xvii


agents’ previous architecture by introducing a GPGP module, which uses the uncoordinated plans from the planner as input, applies appropriate GPGP coordination mechanisms to the uncoordinated plans, and generates coordinated plans to the scheduler for
better schedules.
Fourth, experimentation: applying these implemented coordination mechanisms
to various domains to analyze their performances. Particularly, we chose an emergency
medical service (EMS) domain to demonstrate the effectiveness of the extended set of
GPGP coordination mechanisms. Based on the experimental results, qualitative analysis
has been carried out and significant conclusions have been presented to guide agents in
selecting the best mechanisms in various environments.

xviii


Chapter 1
INTRODUCTION: WHAT AND WHY?

Coordination has been widely studied in many different disciplines, including organization theory, management science, economics, linguistics, psychology, biology, sociology, [64, 81, 83, 84, 99, 146], and of course most importantly to us, computer systems.
Researchers in the area of computer science [43, 53] have opened new vistas in the study
of coordination. Coordination [6, 41, 87, 94, 123, 134] has been studied not only via general or domain-dependent simulation [18, 46, 128, 151], but also through mathematical
formal work [11, 26, 61, 80] in the area of Distributed Artificial Intelligence (DAI). In an
exciting development, coordination technology has been used in many real applications,
such as robot control [3, 110], first response systems [21, 142], web/personal assistant
agents [65, 97], computer games [102], and many more [9, 96, 149].

Coordination is a central research task in distributed artificial intelligence. Well
coordinated behaviors greatly improve intelligent entities’ overall performance. For example, a squad of well coordinated basketball players is more likely to win a game; a
team of emergency rescuers is capable of reaching emergency sites within the shortest
period of time and saving more lives through better coordinated response processes. On
the other hand, inappropriate coordination reduces system efficiency and results in extra complexity; furthermore, failed coordination could leave key tasks unfinished, vital
resources poorly utilized, and even make systems crash. For this reason, the interest in
coordination is constantly growing.

1


1.1 What Is Coordination?
Coordination has different meanings understood by people from diverse areas.
The question of “what is coordination behavior” has been elaborated from multiple view
points: rational systems, natural systems and institutions, open systems, and economic
systems in [44]. Here we only discuss coordination in the area of computer science, particularly as a key topic in distributed artificial intelligence. Definitions of coordination
from various researchers’ points of view are briefly shown as follows.

Composing purposeful actions into larger purposeful wholes.
— Holt, 1983
The property of interaction among some set of agents performing some collective activity.
— Bond and Gasser, 1988
The additional information processing performed when multiple, connected
actors pursue goals that a single actor pursuing the same goals would not
perform.
— Malone, 1988
Activities required to maintain consistency within a work product or to manage dependencies within the work flow
— Curtis, 1989
The act of working together harmoniously
— Malone 1991

The integration and harmonious adjustment of individual work efforts towards the accomplishment of a larger goal
— Singh, 1992
The state of a community of agents in which actions of some agents fit in well
with each other, as well as to the process of achieving this state. The degree of
coordination is the extent to which they avoid extraneous activity by reducing
resource contention, avoiding livelock and deadlock, and maintaining applicable safety conditions
— Weiss, 1999
2


Although these definitions show some more or less similar ideas, each illustrates a particular point-of-view.
The concept that we accept is what Malone and Crowston [98] suggested:
Coordination is managing dependencies between activities.
Malone and Crowston further clarified and upgraded their earlier rather obscure definition,
which was “the act of working together harmoniously.” Obviously the new definition emphasizes the functionality of coordination, which deals with the dependent relationships
among different agents’ tasks. Their definition of coordination suggests that the understanding of coordination demands the characterization of various kinds of dependencies
and features of the coordination processes, which can be also used to manage those dependencies. It is easy to imply that coordination generally happens during the occurrence
of dependencies.
The inter-agent dependencies we will explore in this dissertation can be understood as inter-agent relationships among multiple agents in certain environments, when
the agents take actions to achieve their goals. Several factors affect an agent’s performance in terms of inter-agent relationships. First, agents should choose a course of relevant actions and carefully select the most effective actions to achieve their goals. Second,
agents must pay attention to pending actions in such a sequence that later actions will potentially take advantage of, or at least not be harmed by, the earlier actions. Third, agents
are responsible for taking the right actions at the right times, i.e., a timing issue is also a
key factor to agents’ performance. Finally, no agent is omnipotent or omniscient in the
current complicated multi-agent systems,1 but rather agents hold subjective, or partial,
views of their environments. A single agent can not solve all the problems in the current
complicated systems, maybe not even a very small part of the problems, without the assistance of other agents. The reason lies in either the lack of ability or the lack of resources.
1

Perfect rationality of agent is not possible with bounded computation [13, 67, 82,
119].

3


For example, without cooperating teammates alongside, even the best basketball player
in the world could not win a game alone against a properly coordinated team. This shows
that coordination is needed because of the limited capability of individual agents. For another example, in an emergency medical system (EMS) response process, even the fastest
ambulance could not begin to treat victims still trapped inside burning buildings, unless
they are first rescued by the firefighters; similarly, the ambulance could not transport the
victims to medical facilities in time without police help in severe traffic jams or other
difficult road conditions. The above EMS example implies that coordination is necessary
because of role specificity, another form of the lack of capability. In another case, even
the best cook in the world can not prepare a salad without fruit or vegetables at hand; a
successful meal is based on available raw materials prepared by his assistants. This shows
the importance of resources in coordination processes. Thus, it is easy to envision that
inter-agent relationships are inevitable in complicated multi-agent systems. Within such
systems, if one agent’s action(s) dependents on another agent’s action(s), we call this kind
of relationship a coordination relationship. Further, we name the inter-agent relationship
inter-dependency2 defined as a relationship between a local task (of one agent) and a
non-local task (of another agent) where the execution of one changes some performancerelated characteristics associated with the other, which is based on similar statements
presented in [44]. It explicitly states that interdependency involves tasks across different
agents instead of being limited to the relationship between the tasks of a single agent.
Note that we concentrate on an agent’s (partially) local view. That is, the actions, or
tasks, that can be executed by this agent are called local tasks; the actions or tasks of
other agents are represented locally as non-local tasks. A local task belongs to its owner
agent and can be executed only by its owner agent. From this owner agent’s perspective,
a non-local task involved in a coordination process belongs to a remote agent.
2

Here inter-dependency specifies the inter-agent dependent relationships; the term interdependency without the hyphen will be used throughout this dissertation.

4


Researchers in the field of distributed artificial intelligence presented certain algorithmic methods, called coordination mechanisms, for handling these inter-agent relationships. We define coordination mechanisms as a set of executable procedures and a number
of associated communication protocols for dealing with those coordination relationships.
Different coordination mechanisms may be used in response to different coordination relationships, or sometimes even the same coordination relationships. Many researchers
have shown that there is no single best organization or coordination mechanism for all
environments [44, 54, 56, 66]. Under these assumptions, agents existing as individuals in
distributed computer systems should be able to identify appropriate coordination relationships and adapt their behaviors accordingly to proper coordination mechanisms. Therefore, this dissertation provides solutions to the processes of construction, formalization,
identification, selection, and application of the coordination mechanisms.
1.2 Why study coordination?
In the past decade, agent has become a hot topic and interest in agent technology
has grown dramatically. Agent-related applications are already in practice, and will be
even more popular in the future: a seller’s agent and a buyer’s agent3 are matched with
each other and negotiate the price of a product; information gathering agents roam the Internet to search resources for their owners’ interests; personal assistant agents help us to
communicate with each other by means of email or virtual meetings, discover requested
information in appropriate databases, transfer files back and forth, remind us of everyday
schedules, etc. The wonderful thing about the idea of intelligent agents is that they are
capable of not only understanding what tasks will achieve their goals, but also of reasoning by themselves, and figuring out how to execute the tasks without human intervention.
3

Generally speaking, the agents discussed in this dissertation are always regarded as
computerized software intelligent agents, which are capable of autonomous, flexible,
and potentially social behaviors in order to meet their design objectives, not referring
to human agents.

5


The applications of software intelligent agents will affect our lives deeply.

Currently, most agents are developed to be computationally powerful. However, as
the multi-agent systems grow more and more complicated nowadays, coordination among
multiple intelligent entities is inevitable. The study of the inter-agent relationships needs
to be further explored. As a result, the research of multi-agent coordination becomes
more and more important.
In some previous projects, agents are isolated from each other, trying to optimize
their performance by their own effort, while in most other domains, the agents are not
isolated and can interact with each other in order to achieve their goals by multi-agent
planning, competition, cooperation, negotiation, etc. [78] Thus, agents’ behaviors involve
coordination processes with each other.
Research ideas about coordination have been applied to many applications: the
control of distributed vehicle monitoring, hospital patient scheduling, network meta-level
communication, E-Commerce, Emergent Medical Services (EMS), bioinformatics, RoboRescue, RoboCup, naval radar interference management, military training simulation,
and much more. Some of these applications have been mentioned in the beginning of this
dissertation. More will be briefly introduced later.
Well coordinated behaviors greatly improve intelligent entities’ overall performance. For example, a team of soccer players is more likely to win a game with better
coordination: A mid-back holds the ball at first and at the same time observes the positions of his teammates and the opponents; the strikers and the mid-fielders must run to
find good positions, and get the mid-back’s attention for a feed; the mid-back passes the
ball to the left mid-fielder, who runs to the best position to start the attack; all the other
players move accordingly for their best attacking positions; the left striker approaches the
left mid-fielder to get the ball and kicks a long pass to the right striker; the right striker
shoots for a score. We just described a very successful coordination among a team of soccer players. Effective coordination requires the selection of the appropriate actions, the

6