The only book to present the synergy between modeling and simulation, systems engineering, and agent technologies expands the notion of agent-based simulation to also deal with agent simulation and agent-supported simulation. Accessible to both practitioners and managers, it systematically addresses designing and building agent systems from a systems engineering perspective.
Innehållsförteckning
Foreword VII
Preface XIX
List of Contributors XXIII
Part One Background 1
1 Modeling and Simulation: a Comprehensive and Integrative View 3
Tuncer I. Ören
1.1 Introduction 3
1.2 Simulation: Several Perspectives 4
1.2.1 Purpose of Use 4
1.2.2 Problem to Be Solved 8
1.2.3 Connectivity of Operations 9
1.2.4 M&S as a Type of Knowledge Processing 9
1.2.5 M&S from the Perspective of Philosophy of Science 13
1.3 Model-Based Activities 13
1.3.1 Model Building 15
1.3.2 Model-Base Management 15
1.3.3 Model Processing 15
1.3.4 Behavior Generation 17
1.4 Synergies of M&S: Mutual and Higher-Order Contributions 20
1.5 Advancement of M&S 20
1.6 Preeminence of M&S 24
1.6.1 Physical Tools 27
1.6.2 Knowledge-Based or Soft Tools 27
1.6.3 Knowledge Generation Tools 30
1.7 Summary and Conclusions 32
2 Autonomic Introspective Simulation Systems 37
Levent Yilmaz and Bradley Mitchell
2.1 Introduction 37
2.2 Perspective and Background on Autonomic Systems 39
2.3 Decentralized Autonomic Simulation Systems: Prospects and Issues 41
2.3.1 Motivating Scenario: Adaptive Experience Management in Distributed Mission Training 41
2.3.2 An Architectural Framework for Decentralized Autonomic Simulation Systems 42
2.3.3 Challenges and Issues 44
2.4 Symbiotic Adaptive Multisimulation: An Autonomic Simulation System 47
2.4.1 Metamodels for Introspection Layer Design 48
2.4.2 Local Adaptation: First-Order Change via Particle Swarm Optimizer 50
2.4.3 The Learning Layer: Genetic Search of Potential System Configurations 51
2.4.4 SAMS Component Architecture 52
2.5 Case Study: UAV Search and Attack Scenario 55
2.5.1 Input Factors 56
2.5.2 Agent Specifications 57
2.6 Validation and Preliminary Experimentation with SAMS 64
2.6.1 Face Validity of the UAV Model 65
2.6.2 Experiments with the Parallel SAMS Application 67
2.7 Summary 70
Part Two Agents and Modeling and Simulation 73
3 Agents: Agenthood, Agent Architectures, and Agent Taxonomies 75
Andreas Tolk and Adelinde M. Uhrmacher
3.1 Introduction 75
3.2 Agenthood 76
3.2.1 Defining Agents 76
3.2.2 Situated Environment and Agent Society 78
3.3 Agent Architectures 79
3.3.1 Realizing Situatedness 79
3.3.2 Realizing Autonomy 81
3.3.3 Realizing Flexibility 82
3.3.4 Architectures and Characteristics 84
3.4 Agenthood Implications for Practical Applications 86
3.4.1 Systems Engineering, Simulation, and Agents 87
3.4.2 Modeling and Simulating Human Behavior for Systems Engineering 88
3.4.3 Simulation-Based Testing in Systems Engineering 91
3.4.4 Simulation as Support for Decision Making in Systems Engineering 93
3.4.5 Implications for Modeling and Simulation Methods 94
3.5 Agent Taxonomies 96
3.5.1 History and Application-Specific Taxonomies 96
3.5.2 Categorizing the Agent Space 99
3.6 Concluding Discussion 101
4 Agent-directed Simulation 111
Levent Yilmaz and Tuncer I. Ören
4.1 Introduction 111
4.2 Background 113
4.2.1 Software Agents 113
4.2.2 Complexity 113
4.2.3 Complex Systems of Systems 114
4.2.4 Software Agents within the Spectrum of Computational Paradigms 115
4.3 Categorizing the Use of Agents in Simulation 118
4.3.1 Agent Simulation 118
4.3.2 Agent-Based Simulation 119
4.3.3 Agent-Supported Simulation 119
4.4 Agent Simulation 120
4.4.1 A Metamodel for Agent System Models 120
4.4.2 A Taxonomy for Modeling Agent System Models 122
4.4.3 Using Agents as Model Design Metaphors: Agent-Based Modeling 123
4.4.4 Simulation of Agent Systems 127
4.5 Agent-Based Simulation 129
4.5.1 Autonomic Introspective Simulation 130
4.5.2 Agent-Coordinated Simulator for Exploratory Multisimulation 131
4.6 Agent-Supported Simulation 134
4.6.1 Agent-Mediated Interoperation of Simulations 135
4.6.2 Agent-Supported Simulation for Decision Support 139
4.7 Summary 141
Part Three Systems Engineering and Quality Assurance for Agent-Directed Simulation 145
5 Systems Engineering: Basic Concepts and Life Cycle 147
Steven M. Biemer and Andrew P. Sage
5.1 Introduction 147
5.2 Agent-Based Systems Engineering 148
5.3 Systems Engineering Definition and Attributes 148
5.3.1 Knowledge 149
5.3.2 People and Information Management 150
5.3.3 Processes 151
5.3.4 Methods and Tools 156
5.3.5 The Need for Systems Engineering 157
5.4 The System Life Cycle 157
5.4.1 Conceptual Design (Requirements Analysis) 160
5.4.2 Preliminary Design (Systems Architecting) 161
5.4.3 Detailed Design and Development 161
5.4.4 Production and Construction 163
5.4.5 Operational Use and System Support 164
5.5 Key Concepts of Systems Engineering 164
5.5.1 Integrating Perspectives into the Whole 164
5.5.2 Risk Management 165
5.5.3 Decisions and Trade Studies (the Strength of Alternatives) 166
5.5.4 Modeling and Evaluating the System 168
5.6 Summary 169
6 Quality Assurance of Simulation Studies of Complex Networked Agent Systems 173
Osman Balci, William F. Ormsby, and Levent Yilmaz
6.1 Introduction 173
6.2 Characteristics of Open Agent Systems 174
6.3 Issues in the Quality Assurance of Agent Simulations 175
6.4 Large-Scale Open Complex Systems – The Network-Centric System Metaphor 177
6.5 M&S Challenges for Large-Scale Open Complex Systems 179
6.6 Quality Assessment of Simulations of Large-Scale Open Systems 181
6.7 Conclusions 186
7 Failure Avoidance in Agent-directed Simulation: Beyond Conventional v&v and qa 189
Tuncer I. Ören and Levent Yilmaz
7.1 Introduction 189
7.1.1 The Need for a Fresh Look 189
7.1.2 Basic Terms 191
7.2 What Can Go Wrong 192
7.2.1 Increasing Importance of M&S 192
7.2.2 Contributions of Simulation to Failure Avoidance 192
7.2.3 Need for Failure Avoidance in Simulation Studies 194
7.2.4 Some Sources of Failure in M&S 196
7.3 Assessment for M&S 198
7.3.1 Types of Assessment 198
7.3.2 Criteria for Assessment 200
7.3.3 Elements of M&S to be Studied 200
7.4 Need for Multiparadigm Approach for Successful M&S Projects 200
7.4.1 V&V Paradigm for Successful M&S Projects 201
7.4.2 QA Paradigm for Successful M&S Projects 203
7.4.3 Failure Avoidance Paradigm for Successful M&S Projects 204
7.4.4 Lessons Learned and Best Practices for Successful M&S Projects 204
7.5 Failure Avoidance for Agent-Based Modeling 206
7.5.1 Failure Avoidance in Rule-Based Systems 207
7.5.2 Failure Avoidance in Autonomous Systems 208
7.5.3 Failure Avoidance in Agents with Personality, Emotions, and Cultural Background 209
7.5.4 Failure Avoidance in Inputs 210
7.6 Failure Avoidance for Systems Engineering 212
7.7 Conclusion 213
8 Toward Systems Engineering for Agent-directed Simulation 219
Levent Yilmaz
8.1 Introduction 219
8.2 What Is a System? 220
8.2.1 What Is Systems Engineering? 220
8.2.2 The Functions of Systems Engineering 220
8.3 Modeling and Simulation 221
8.4 The Synergy of M&S and SE 221
8.4.1 The Role of M&S in Systems 221
8.4.2 Why Does M&S Require SE? 222
8.4.3 Why Is SSE Necessary? 222
8.5 Toward Systems Engineering for Agent-Directed Simulation 222
8.5.1 The Essence of Complex Adaptive Open Systems (CAOS) 223
8.5.2 The Merits of ADS 224
8.5.3 Systems Engineering for Agent-Directed Simulation 225
8.6 Sociocognitive Framework for ADS-SE 225
8.6.1 Social-Cognitive View 226
8.6.2 The Dimensions of Representation 227
8.6.3 The Functions for Analysis 228
8.7 Case Study: Human-Centered Work Systems 228
8.7.1 Operational Level – Organizational Subsystem 229
8.7.2 Operational Level – Organizational Subsystem 230
8.7.3 Operational Level – Integration of Organization and Social Subsystems 232
8.7.4 The Technical Level 232
8.8 Conclusions 235
9 Design and Analysis of Organization Adaptation in Agent Systems 237
Virginia Dignum, Frank Dignum, and Liz Sonenberg
9.1 Introduction 237
9.2 Organizational Model 239
9.3 Organizational Structure 240
9.3.1 Organizational Structures in Organization Theory 240
9.3.2 Organizational Structures in Multiagent Systems 241
9.4 Organization and Environment 242
9.4.1 Environment Characteristics 242
9.4.2 Congruence 244
9.5 Organization and Autonomy 245
9.6 Reorganization 247
9.6.1 Organizational Utility 247
9.6.2 Organizational Change 248
9.7 Organizational Design 250
9.7.1 Designing Organizational Simulations 252
9.7.2 Application Scenario 253
9.8 Understanding Simulation of Reorganization 256
9.8.1 Reorganization Dimensions 257
9.8.2 Analyzing Simulation Case Studies 257
9.9 Conclusions 263
10 Programming Languages, Environments, and Tools for Agent-directed Simulation 269
Yu Zhang, Mark Lewis, and Maarten Sierhuis
10.1 Introduction 269
10.2 Architectural Style for ADS 271
10.3 Agent-Directed Simulation – An Overview 272
10.3.1 Language 273
10.3.2 Environment 275
10.3.3 Service 276
10.3.4 Application 276
10.4 A Survey of Five ADS Platforms 277
10.4.1 Ascape 277
10.4.2 Net Logo 280
10.4.3 Repast 283
10.4.4 Swarm 286
10.4.5 Mason 289
10.5 Brahms – A Multiagent Simulation for Work System Analysis and Design 291
10.5.1 Language 291
10.5.2 Environment 295
10.5.3 Service 298
10.5.4 Application 299
10.6 CASESim – A Multiagent Simulation for Cognitive Agents for Social Environment 300
10.6.1 Language 302
10.6.2 Environment 302
10.6.3 Service 306
10.6.4 Application 310
10.7 Conclusion 312
11 Simulation for Systems Engineering 317
Joachim Fuchs
11.1 Introduction 317
11.2 The Systems Engineering Process 317
11.3 Modeling and Simulation Support 318
11.4 Facilities 320
11.5 An Industrial Use Case: Space Systems 321
11.5.1 Simulators for Analysis and Design 323
11.5.2 Facility for Spacecraft Qualification and Acceptance 325
11.5.3 Facility for Ground System Qualification and Testing and Operations 325
11.6 Outlook 325
11.7 Conclusions 327
12 Agent-directed Simulation for Systems Engineering 329
Philip S. Barry, Matthew T.K. Koehler, and Brian F. Tivnan
12.1 Introduction 329
12.2 New Approaches Are Needed 331
12.2.1 Employing ADS Through the Framework of Empirical Relevance 332
12.2.2 Simulating Systems of Systems 334
12.3 Agent-Directed Simulation for the Systems Engineering of Human Complex Systems 336
12.3.1 A Call for Agents in the Study of Human Complex Systems 337
12.3.2 Noteworthy Agent-Directed Simulations in the Science of Human Complex Systems 338
12.4 A Model-Centered Science of Human Complex Systems 338
12.5 An Infrastructure for the Engineering of Human Complex Systems 339
12.5.1 Components of the Infrastructure for Complex Systems Engineering 339
12.5.2 Modeling Goodness 341
12.5.3 The Genetic Algorithm Optimization Toolkit 341
12.6 Case Studies 344
12.6.1 Case Study 1: Defending The Stadium 345
12.6.2 Case Study 2: Secondary Effects from Pandemic Influenza 350
12.7 Summary 355
Part Four Agent-Directed Simulation for Systems Engineering 361
13 Agent-implemented Experimental Frames for Net-centric Systems Test and Evaluation 363
Bernard P. Zeigler, Dane Hall, and Manuel Salas
13.1 Introduction 363
13.2 The Need for Verification Requirements 364
13.3 Experimental Frames and System Entity Structures 366
13.4 Decomposition and Design of System Architecture 371
13.5 Employing Agents in M&S-Based Design, Verification and Validation 376
13.6 Experimental Frame Concepts for Agent Implementation 378
13.7 Agent-Implemented Experimental Frames 381
13.8 DEVS/SOA: Net-Centric Execution Using Simulation Service 382
13.8.1 Automation of Agent Attachment to System Components 382
13.8.2 DEVS-Agent Communications/Coordination 384
13.8.3 DEVS-Agent Endomorphic Models 386
13.9 Summary and Conclusions 388
13.A c Auto DEVS – A Tool for the Bifurcated Methodology 391
14 Agents and Decision Support Systems 399
Andreas Tolk, Poornima Madhavan, Jeffrey W. Tweedale, and Lakhmi C. Jain
14.1 Introduction 399
14.1.1 History 399
14.1.2 Motivating Agent-Directed Decision Support Simulation Systems 401
14.1.3 Working Definitions 403
14.2 Cognitive Foundations for Decision Support 405
14.2.1 Decision Support Systems as Social Actors 406
14.2.2 How to Present the System to the User and Improve Trust 407
14.2.3 Relevance for the Engineer 410
14.3 Technical Foundations for Decision Support 411
14.3.1 Machine-Based Understanding for Decision Support 412
14.3.2 Requirements for Systems When Being Used for Decision Support 413
14.3.3 Agent-Directed Multimodel and Multisimulation Support 417
14.3.4 Methods Applicable to Support Agent-Directed Decision Support Simulation Systems 418
14.4 Examples for Intelligent and Agent-Directed Decision Support Simulation Systems 421
14.4.1 Supporting Command and Control 421
14.4.2 Supporting Inventory Control and Integrated Logistics 423
14.5 Conclusion 426
15 Agent Simulation for Software Process Performance Analysis 433
Levent Yilmaz and Jared Phillips
15.1 Introduction 433
15.2 Related Work 435
15.2.1 Organization-Theoretic Perspective for Simulation-Based Analysis of Software Processes 435
15.2.2 Simulation Methods for Software Process Performance Analysis 436
15.3 Team-RUP: A Framework for Agent Simulation of Software Development Organizations 437
15.3.1 Organization Structure 437
15.3.2 Team-RUP Task Model 438
15.3.3 Team-RUP Team Archetypes and Cooperation Mechanisms 439
15.3.4 Reward Mechanism in Team-RUP 440
15.4 Design and Implementation of Team-RUP 441
15.4.1 Performance Metrics 443
15.4.2 Validation of the Model 444
15.5 Results and Discussion 445
15.6 Conclusions 447
16 Agent-Directed Simulation for Manufacturing System Engineering 451
Jeffrey S. Smith, Erdal Sahin, and Levent Yilmaz
16.1 Introduction 451
16.1.1 Manufacturing Systems 452
16.1.2 Agent-Based Modeling 453
16.2 Simulation Modeling and Analysis for Manufacturing Systems 454
16.2.1 Manufacturing System Design 455
16.2.2 Manufacturing Operation 458
16.3 Agent-Directed Simulation for Manufacturing Systems 463
16.3.1 Emergent Approaches 463
16.3.2 Agent-Based Manufacturing 464
16.3.3 The Holonic Approach: Hierarchic Open Agent Systems 466
16.4 Summary 468
17 Organization and Work Systems Design and Engineering: from Simulation to Implementation of Multiagent Systems 475
Maarten Sierhuis, William J. Clancey, and Chin H. Seah
17.1 Introduction 475
17.2 Work Systems Design 475
17.2.1 Existing Work System Design Methods 476
17.2.2 A Brief History of Work Systems Design 477
17.3 Modeling and Simulation of Work Systems 478
17.3.1 Designing Work Systems: What Is the Purpose and What Can Go Wrong? 478
17.3.2 The Difficulty of Convincing Management 479
17.4 Work Practice Modeling and Simulation 480
17.4.1 Practice vs. Process 481
17.4.2 Modeling Work Practice 481
17.5 The Brahms Language 487
17.5.1 Simulation or Execution with Brahms 488
17.5.2 Modeling People and Organizations 489
17.5.3 Modeling Artifacts and Data Objects 490
17.5.4 Modeling Communication 492
17.5.5 Modeling Location and Movement 493
17.5.6 Java Integration 495
17.6 Systems Engineering: From Simulation to Implementation 496
17.6.1 A Cyclic Approach 498
17.6.2 Modeling Current Operations 499
17.6.3 Modeling Future Operations 501
17.6.4 MAS Implementation 502
17.7 A Case Study: The OCA Mirroring System 503
17.7.1 Mission Control as a Socio-Technical Work System 504
17.7.2 The OCA Officer’s Work System 505
17.7.3 Simulating the Current OCA Work System 505
17.7.4 Designing the Future OCA Work System 510
17.7.5 Simulating the Future OCA Work System 511
17.7.6 Implementing OCAMS 511
17.8 Conclusion 514
Index 517
Om författaren
Levent Yilmaz is assistant professor of computer science and software engineering at the College of Engineering, Auburn University, USA. Before joining the faculty in 2003, Professor Yilmaz worked as a senior research engineer in the Simulation and Software Division of Trident Systems, Inc., where he held the position of a lead project engineer and principle investigator for advanced simulation methodology, model-based verification, and simulation interoperability efforts. Professor Yilmaz received his Ph.D. and M.S. degrees from the Virginia Polytechnic Institute and State University, Blacksburg, USA.
Tuncer I. Oren is professor emeritus of computer science at the School of Information Technology and Engineering (SITE) of the University of Ottawa, Canada, where he held a chair as full professor from 1981 to 1996. Professor Oren’s research interests focus on the topics of modelling and simulation, agent-directed simulation, cognitive simulation, reliability and quality, and ethics in simulation. He has published over 300 papers and several books.