Artificial war : multiagent-based simulation of combat / Andrew Ilachinski.

Format:

Book

Author/Creator:

Ilachinski, Andrew.

Language:

English

Subjects (All):

War--Mathematical models.

War.

War--Computer simulation.

Physical Description:

xxxiii, 747 pages : illustrations (some color) ; 26 cm

Place of Publication:

River Edge, NJ : World Scientific Pub., [2004]

Summary:

Military conflicts, particularly land combat, possess the characteristics of complex adaptive systems: combat forces are composed of a large number of nonlinearly interacting parts and are organized in a dynamic command-and-control network; local action, which often appears disordered, self-organizes into long-range order; military conflicts, by their nature, proceed far from equilibrium; military forces adapt to a changing combat environment; and there is no master "voice" that dictates the actions of every soldier (i.e., battlefield action is decentralized). Nonetheless, most modern "state of the art" military simulations ignore the self-organizing properties of combat. This book summarizes the results of a multiyear research effort aimed at exploring the applicability of complex adaptive systems theory to the study of warfare, and introduces a sophisticated multiagent-based simulation of combat called EINSTein. EINSTein, whose bottom-up, generative approach to modeling combat stands in stark contrast to the top-down or reductionist philosophy that still underlies most conventional military models, is designed to illustrate how many aspects of land combat may be understood as self-organized, emergent phenomena. Used worldwide by the military operations research community, EINSTein has pioneered the simulation of combat on a small to medium scale by using autonomous agents to model individual behaviors and personalities rather than hardware.

Contents:

1.1 Brief History of CNA's Complexity & Combat Research Project 2

1.1.1 The "Problem" 2

1.1.2 Applying the "New Sciences" to Warfare 7

1.1.3 Warfare & Complexity 12

1.1.4 ISAAC 14

1.1.5 EINSTein 17

1.2 Background and Motivations 22

1.2.1 Lanchester Equations of Combat 22

1.2.2 Artificial Life 25

1.3 Models & Simulations: A Heuristic Discussion 29

1.3.2 Connection to Reality 32

1.3.3 Mathematical Models 35

1.3.4 Computer Simulations 36

1.3.5 What Price Complexity? 37

1.4 Combat Simulation 39

1.4.1 Modeling and Simulation Master Plan 40

1.4.2 Modeling Human Behavior and Command Decision-Making 41

1.4.3 Conventional Simulations 41

1.4.4 Future of Modeling Technology 43

1.5 Multiagent-Based Models and Simulations 44

1.5.1 Autonomous Agents 45

1.5.2 How is Multiagent-Based Modeling Really Done? 46

1.5.3 Agent-Based Simulations vs. Traditional Mathematical Models 48

1.5.4 Multiagent-Based Simulations vs. Traditional AI 50

1.5.5 Examples of MultiAgent-Based Simulations 51

1.5.6 Value of Multiagent-Based Simulations 53

1.5.7 CA-Based & Other EINSTein-Related Combat Models 55

1.6 EINSTein as an Exemplar of More General Models of Complex Adaptive Systems 59

1.6.1 Persian Gulf Scenario 60

1.6.2 SCUDHunt 60

1.6.3 Social Modeling: Riots and Civil Unrest 62

1.6.4 General Applications 63

1.6.5 Universal Patterns of Behavior 64

1.7 Goals & Payoffs for Developing EINSTein 65

1.7.1 Command & Control 65

1.7.2 Pattern Recognition 66

1.7.3 "What If?" Experimentation 66

1.7.4 Fundamental Grammar of Combat? 67

1.8 Toward an Axiological Ontology of Complex Systems 67

1.8.1 Why "Value"? 67

1.8.2 Why "Axiological Ontology"? 68

Chapter 2 Nonlinear Dynamics, Deterministic Chaos and Complex Adaptive Systems: A Primer 71

2.1 Nonlinear Dynamics and Chaos 72

2.1.1 Brief History 72

2.1.2 Dynamical Systems 74

2.1.3 Deterministic Chaos 77

2.1.4 Qualitative Characterization of Chaos 90

2.1.5 Quantitative Characterization of Chaos 92

2.1.6 Time-Series Forecasting and Predictability 98

2.1.7 Chaotic Control 100

2.2 Complex Adaptive Systems 101

2.2.2 Short History 103

2.2.3 General Properties: A Heuristic Discussion 105

2.2.4 Measures of Complexity 114

2.2.5 Complexity as Science: Toward a New Worldview? 129

2.2.6 Artificial Life 131

2.2.7 Cellular Automata 137

2.2.8 Self-Organized Criticality 149

Chapter 3 Nonlinearity, Complexity, and Warfare: Eight Tiers of Applicability 159

3.1 Approach 160

3.2 Tier I: General Metaphors for Complexity in War 162

3.2.1 What is a Metaphor? 162

3.2.2 Metaphors and War 164

3.2.3 Metaphor Shift 166

3.3 Tier II: Policy and General Guidelines for Strategy 171

3.3.1 What Does the New Metaphor Give Us? 171

3.3.2 Policy 171

3.3.3 Organizational Structure 173

3.3.4 Intelligence Analysis 173

3.3.5 Policy Exploitation of Characteristic Time Scales of Combat 174

3.4 Tier III: "Conventional" Warfare Models and Approahces 175

3.4.1 Testing for the Veracity of Conventional Models 176

3.4.2 Non-Monoticities and Chaos 177

3.4.3 Minimalist Modeling 178

3.4.4 Generalizations of Lanchester's equations 179

3.4.5 Nonlinear Dynamics and Chaos in Arms-Race Models 181

3.5 Tier IV: Description of the Complexity of Combat 182

3.5.1 Attractor Reconstruction from Time-Series Data 182

3.5.2 Fractals and Combat 183

3.5.3 Evidence of Chaos in War From Historical Data? 184

3.5.4 Evidence of Self-Organized Criticality From Historical Data? 185

3.5.5 Use of Complex Systems Inspired Measures to Describe Combat 186

3.5.6 Use of Relativistic Information to Describe Command and Control Processes 189

3.6 Tier V: Combat Technology Enhancement 190

3.6.1 Computer Viruses ("computer counter-measures") 190

3.6.2 Fractal Image Compression 190

3.6.3 Cryptography 192

3.7 Tier VI: Combat Aids 193

3.7.1 Using Genetic Algorithms to Evolve Tank Strategies 194

3.7.2 Tactical Decision Aids 197

3.7.3 Classifier Systems 199

3.7.4 How can Genetic Algorithms be Used? 200

3.7.5 Tactical Picture Agents 201

3.8 Tier VII: Synthetic Combat Environments 202

3.8.1 Combat Simulation using Cellular Automata 202

3.8.2 Multiagent-Based Simulations 204

3.9 Tier VIII: Original Conceptualizations of Combat 204

3.9.1 Dueling Parasites 205

3.9.2 Percolation Theory and Command and Control Processes 206

3.9.3 Exploiting Chaos 207

3.9.4 Pattern Recognition 209

3.9.5 Fire-Ant Warfare 214

Chapter 4 EINSTein: Mathematical Overview 217

4.2 Design Philosophy 220

4.2.1 Agent Hierarchy 221

4.2.2 Guiding Principles 222

4.3 Abstract Agent Architecture 224

4.3.2 Dynamics of Value 226

4.3.3 General Formalism 226

4.3.4 Agents in EINSTein 228

4.3.5 Actions 230

4.3.6 Features 234

4.3.7 Local Context 236

4.3.9 Ontological Partitioning 239

4.3.10 Communication 240

4.3.11 Axiological Ontology 241

4.3.12 Preventing a Combinatorial Explosion 242

Chapter 5 EINSTein: Methodology 277

5.1 Program Structure 277

5.1.1 Source Code 277

5.1.2 Object-Oriented 278

5.1.3 Program Flow 280

5.2 Combat Engine 281

5.2.1 Agents 281

5.2.2 Battlefield 283

5.2.3 Agent Sensor Parameters 284

5.2.4 Agent Personalities 286

5.2.5 Agent Action Selection 287

5.2.6 Move Decision Logic Flags 298

5.2.7 Meta-Rules 298

5.2.8 Decision Logic 310

5.2.9 Ambiguity Resolution Logic 311

5.3 Squads 312

5.3.1 Inter-Squad Weight Matrix 313

5.4 Combat 313

5.4.1 As Implemented in Versions 1.0 and Earlier 314

5.4.2 As Implemented in Versions 1.1 and Later 323

5.5 Communications 334

5.5.1 Inter-Squad Communication Weight Matrix 337

5.6 Terrain 337

5.6.1 As Implemented in Versions 1.0 and Earlier 338

5.6.2 As Implemented in Versions 1.1 and Newer 340

5.7 Finding and Navigating Paths 342

5.7.1 Pathfinding 343

5.7.2 Navigating User-Defined Paths 348

5.8 Command and Control 360

5.8.1 Local Command 362

5.8.2 Subordinate Agents 364

5.8.4 Global Command 365

Technical Appendix 1 Enhanced Action Selection Logic 371

Technical Appendix 2 Trigger State Activation 380

Technical Appendix 3 FindWeights 386

Technical Appendix 4 Weight Modification via Internal Feature Space 400

Technical Appendix 5 Action Logic Function (ALF) 403

Technical Appendix 6 Previsualizing Agent Behaviors 408

Chapter 6 EINSTein: Sample Behavior 433

6.1.1 Simulation Run Modes 434

6.1.2 Observations 435

6.1.3 Classes of Behavior 436

6.2 Case Study 1: Lanchesterian Combat 438

6.3 Case Study 2: Classic Battle Front (Tutorial) 441

6.3.1 Collecting Data 441

6.3.2 Asking "What If?" Questions 442

6.3.3 Generating a Fitness Landscape 450

6.4 Case Study 3: Explosive Skirmish 453

6.4.1 Agent-Density Plots 454

6.4.2 Spatial Entropy 454

6.4.3 Fractal Dimensions and Combat 457

6.4.4 Attrition Count 462

6.4.5 Attrition Rate 466

6.5 Case Study 4: Squad vs.

Squad 470

6.5.2 Scenario Definition 471

6.5.3 Weapon Scaling 472

6.5.4 3:1 Force Ratio Rule-of-Thumb 474

6.6 Case Study 5: Attack 476

6.7 Case Study 6: Defense 479

6.8 Case Study 7: Swarms 482

6.9 Case Study 8: Non-Monotonicity 483

6.10 Case Study 9: Autopoietic Skirmish 487

6.11 Case Study 10: Small Insertion 488

6.12 Case Study #11: Miscellaneous Behaviors 492

6.12.1 Precessional Maneuver 492

6.12.2 Random Defense 494

6.12.3 Communications 496

6.12.4 Local Command 497

6.12.5 Global Command 498

Chapter 7 Breeding Agents 501

7.1.1 Genetic Operators 502

7.1.2 The Fitness Landscape 503

7.1.3 The Basic GA Recipe 506

7.1.4 How Do GAs Work? 508

7.2 GAs Adapted to EINSTein 510

7.2.1 Mission Fitness Measures 512

7.2.2 Fitness Function 513

7.2.3 EINSTein's GA Recipe 520

7.2.4 EINSTein's GA Search Space 521

7.3 GA Breeding Experiments 525

7.3.1 Agent "Breeding" Experiment #1 (Tutorial) 525

7.3.2 Agent "Breeding" Experiment #2 534

7.3.3 Agent "Breeding" Experiment #3 537

7.3.4 Agent "Breeding" Experiment #4 539

Chapter 8 Concluding Remarks & Speculations 543

8.1 EINSTein 545

8.3 Payoffs 550

8.4 Validation 551

8.4.1 EINSTein and JANUS 553

8.4.2 Alignment of Computational models 554

8.5 Future Work 555

8.6 Final Comment 560

A.2 Adaptive Systems 561

A.3 Agents 562

A.4 Artificial Intelligence 562

A.5 Artificial Life 562

A.6 Cellular Automata 563

A.7 Chaos 563

A.8 Complexity 564

A.9 Conflict & War 564

A.10 Fuzzy Logic 565

A.11 Game Programming 565

A.12 Genetic Algorithms 565

A.13 Information Visualization 565

A.14 Machine Learning 566

A.15 Newsgroups 566

A.16 Philosophical 566

A.17 Robotics 567

A.18 Simulation Systems 567

A.19 Swarm Intelligence 568

A.20 Time Series Analysis 568

Appendix B EINSTein Homepage 569

B.1 Links 569

B.2 Screenshots 570

Appendix C EINSTein Development Tools 573

Appendix D Installing EINSTein 575

D.1 Versions 575

D.2 System Requirements 575

D.3 Installing EINSTein 575

D.4 Running EINSTein 576

Appendix E A Concise User's Guide to EINSTein 581

E.1 File Menu 581

E.1.1 Load... 581

E.1.2 Save... 583

E.1.3 Exit 584

E.2 Edit Menu 584

E.2.1 Combat Parameters... 584

E.2.2 Red Data 586

E.2.3 Terrain 598

E.2.4 Territorial Possession 600

E.2.5 Multiple Time-Series Run Parameters 601

E.2.6 2-Parameter Fitness Landscape Exploration 602

E.2.7 1-Sided Genetic Algorithm Parameters 602

E.3 Simulation Menu 603

E.3.1 Interactive Run Mode 603

E.3.2 Play-Back Run Mode 604

E.3.3 Multiple Time-Series Run Mode 604

E.3.4 2-Parameter Phase Space Exploration 607

E.3.5 One-Sided Genetic Algorithm Run Mode 609

E.3.6 Clear 612

E.3.7 Run/Stop Toggle 612

E.3.8 Step-Execute Mode 613

E.3.9 Step Execute for T Steps... 613

E.3.10 Randomize 613

E.3.11 Reseed Random Number Generator 613

E.3.12 Restart... 614

E.3.13 Terminate Run 615

E.4 Display Menu 615

E.4.1 Data 616

E.4.2 Toggle Background Color 618

E.4.3 Trace Map 618

E.4.4 Display All Agents (Default) 618

E.4.5 Display All Agents (Highlight Injured) 619

E.4.6 Display Alive Agents Alone 619

E.4.7 Display Injured Agents Alone 619

E.4.8 Highlight Individual Squad 619

E.4.9 Highlight Command Structure 620

E.4.10 Activity Map 621

E.4.11 Battle-Front Map 622

E.4.12 Killing Field Map 623

E.4.13 Territorial Possession Map 624

E.4.14 Zoom 626

E.5 On-the-Fly Parameter Changes Menu 626

E.5.1 EINSTein's On-the-Fly Parameter Changes Menu Options 627

E.6 Data Collection Menu 628

E.6.1 Toggle Data Collection On/Off 628

E.6.2 Set All 628

E.6.3 Capacity Dimension 628

E.6.4 Force Sizes 629

E.6.5 Center-of-Mass Positions 629

E.6.6 Cluster-Size Distributions 629

E.6.7 Goal Count 629

E.6.8 Interpoint Distance Distributions 629

E.6.9 Neighbor-Number Distributions 630

E.6.10 Spatial Entropy 630

E.6.11 Territorial Possession 630

E.6.12 Mission-Fitness Landscape (2-Parameter)... 630

E.6.13 Calculate Capacity Dimension (Snapshot at time t) 631

E.7 Data Visualization Menu 631

E.7.1 2D Graphs 632

E.7.2 3D Graphs 643

E.8 Help Menu 645

E.8.1 Help Topics 645

E.8.2 About EINSTein... 646

E.9 Toolbar 647

E.9.1 Toolbar Reference 647

Appendix F Differences Between EINSTein Versions 1.0 (and older) and 1.1 (and newer) 651

F.1 Toolbar and Main Menu 651

F.2 Main Menu Edit Options and Dialogs 651

F.2.1 Agent Paraemters 652

F.2.2 Edit Terrain Type 654

F.2.3 Combat-Related Dialogs 655

F.2.4 Main Menu Simulation Pptions/Dialogs 656

F.2.5 Main Menu Display Options 657

F.2.6 Right-Hand Mouse Action 658

Appendix G EINSTein's Data Files 663

G.1 Versions 1.0 and Earlier 663

G.1.1 Input Data File 664

G.1.2 Combat Agent Input Data File 689

G.1.3 Run-File 690

G.1.4 Terrain Input Data File 690

G.1.5 Terrain-Modified Agent Parameters Input Data File 691

G.1.6 Weapons Input Data File 692

G.1.7 Two-Parameter Fitness Landscape Input Data File 693

G.1.8 One-Sided Genetic Algorithm Input Data File 697

G.1.9 Communications Matrix Input Data File 700

G.1.10 Squad Interconnectivity Matrix Input Data File 701

G.1.11 Output Data Files 701

G.2 Versions 1.1 and Newer 707.

Notes:

Includes bibliographical references (pages 709-738) and index.

ISBN:

9812388346

OCLC:

55473560