Probabilistic reasoning in multiagent systems : a graphical models approach / Yang Xiang.

Format:

Book

Author/Creator:

Xiang, Yang, 1954-

Contributor:

Benjamin Franklin Library Fund.

Language:

English

Subjects (All):

Distributed artificial intelligence.

Bayesian statistical decision theory--Data processing.

Bayesian statistical decision theory.

Intelligent agents (Computer software).

Physical Description:

xii, 294 pages : illustrations ; 26 cm

Place of Publication:

Cambridge ; New York : Cambridge University Press, 2002.

Summary:

Addresses the challenges of building intelligent agents to cooperate on complex tasks in uncertain environments.

Contents:

1.1 Intelligent Agents 1

1.2 Reasoning about the Environment 4

1.3 Why Uncertain Reasoning? 5

1.4 Multiagent Systems 7

1.5 Cooperative Multiagent Probabilistic Reasoning 11

1.6 Application Domains 13

2 Bayesian Networks 16

2.2 Basics on Bayesian Probability Theory 19

2.3 Belief Updating Using JPD 23

2.4 Graphs 24

2.6 Local Computation and Message Passing 30

2.7 Message Passing over Multiple Networks 31

2.8 Approximation with Massive Message Passing 33

3 Belief Updating and Cluster Graphs 37

3.3 Conventions for Message Passing in Cluster Graphs 43

3.4 Relation with [lambda] - [pi] Message Passing 44

3.5 Message Passing in Nondegenerate Cycles 47

3.6 Message Passing in Degenerate Cycles 53

3.7 Junction Trees 56

4 Junction Tree Representation 61

4.2 Graphical Separation 64

4.3 Sufficient Message and Independence 68

4.4 Encoding Independence in Graphs 69

4.5 Junction Trees and Chordal Graphs 71

4.6 Triangulation by Elimination 76

4.7 Junction Trees as I-maps 78

4.8 Junction Tree Construction 80

5 Belief Updating with Junction Trees 86

5.2 Algebraic Properties of Potentials 88

5.3 Potential Assignment in Junction Trees 94

5.4 Passing Belief over Separators 97

5.5 Passing Belief through a Junction Tree 100

5.6 Processing Observations 104

6 Multiply Sectioned Bayesian Networks 107

6.2 The Task of Distributed Uncertain Reasoning 112

6.3 Organization of Agents during Communication 117

6.4 Agent Interface 124

6.5 Multiagent Dependence Structure 128

7 Linked Junction Forests 142

7.2 Multiagent Distributed Compilation of MSBNs 146

7.3 Multiagent Moralization of MSDAG 147

7.4 Effective Communication Using Linkage Trees 152

7.5 Linkage Trees as I-maps 155

7.6 Multiagent Triangulation 158

7.7 Constructing Local Junction Trees and Linkage Trees 174

8 Distributed Multiagent Inference 182

8.2 Potentials in a Linked Junction Forest 186

8.3 Linkage Trees over the Same d-sepset 190

8.4 Extended Linkage Potential 192

8.5 E-message Passing between Agents 194

8.6 Multiagent Communication 196

8.7 Troubleshooting a Digital System 201

8.8 Complexity of Multiagent Communication 207

8.9 Regional Multiagent Communication 208

8.10 Alternative Methods for Multiagent Inference 209

9 Model Construction and Verification 215

9.2 Multiagent MSBN System Integration 217

9.3 Verification of Subdomain Division 219

9.4 Agent Registration 221

9.5 Distributed Verification of Acyclicity 223

9.6 Verification of Agent Interface 237

9.7 Complexity of Cooperative d-sepset Testing 271

10 Looking into the Future 274

10.1 Multiagent Reasoning in Dynamic Domains 274

10.2 Multiagent Decision Making 277

10.3 What If Verification Fails? 279

10.4 Dynamic Formation of MSBNs 279

10.5 Knowledge Adaptation and Learning 280

10.6 Negotiation over Agent Interfaces 281

10.7 Relaxing Hypertree Organization 283

10.8 Model Approximation 284

10.9 Mixed Models 285.

Notes:

Includes bibliographical references (pages 287-291) and index.

Local Notes:

Acquired for the Penn Libraries with assistance from the Benjamin Franklin Library Fund.

ISBN:

0521813085

OCLC:

48449503

Online:

Publisher description