Swarm Intelligence

by Eberhart; Shi; Kennedy

• ISBN13: 9781558605954
• ISBN10: 1558605959
• Format: Hardcover
• Copyright: 2001-03-26
• Publisher: Elsevier Science

Summary

Traditional methods for creating intelligent computational systems have privileged private "internal" cognitive and computational processes. In contrast, Swarm Intelligence argues that human intelligence derives from the interactions of individuals in a social world and further, that this model of intelligence can be effectively applied to artificially intelligent systems. The authors first present the foundations of this new approach through an extensive review of the critical literature in social psychology, cognitive science, and evolutionary computation. They then show in detail how these theories and models apply to a new computational intelligence methodologyparticle swarmswhich focuses on adaptation as the key behavior of intelligent systems. Drilling down still further, the authors describe the practical benefits of applying particle swarm optimization to a range of engineering problems. Developed by the authors, this algorithm is an extension of cellular automata and provides a powerful optimization, learning, and problem solving method. This important book presents valuable new insights by exploring the boundaries shared by cognitive science, social psychology, artificial life, artificial intelligence, and evolutionary computation and by applying these insights to the solving of difficult engineering problems. Researchers and graduate students in any of these disciplines will find the material intriguing, provocative, and revealing as will the curious and savvy computing professional. * Places particle swarms within the larger context of intelligent adaptive behavior and evolutionary computation. * Describes recent results of experiments with the particle swarm optimization (PSO) algorithm * Includes a basic overview of statistics to ensure readers can properly analyze the results of their own experiments using the algorithm. * Support software which can be downloaded from the publishers website, includes a Java PSO applet, C and Visual Basic source code.

Author Biography

James Kennedy is a social psychologist who works in survey methods at the U.S. Department of Labor Russell C. Eberhart is Associate Dean for Research, Purdue School of Engineering and Technology, Indiana University Purdue University Indianapolis Yuhui Shi received his Ph.D. in electrical engineering from Southeast University, China, in 1992 and is an applied specialist for Electronic Data Systems, Inc.

Table of Contents

Preface	p. xiii
Foundations
Models and Concepts of Life and Intelligence	p. 3
The Mechanics of Life and Thought	p. 4
Stochastic Adaptation: Is Anything Ever Really Random?	p. 9
The "Two Great Stochastic Systems"	p. 12
The Game of Life: Emergence in Complex Systems	p. 16
The Game of Life	p. 17
Emergence	p. 18
Cellular Automata and the Edge of Chaos	p. 20
Artificial Life in Computer Programs	p. 26
Intelligence: Good Minds in People and Machines	p. 30
Intelligence in People: The Boring Criterion	p. 30
Intelligence in Machines: The Turing Criterion	p. 32
Symbols, Connections, and Optimization by Trial and Error	p. 35
Symbols in Trees and Networks	p. 36
Problem Solving and Optimization	p. 48
A Super-Simple Optimization Problem	p. 49
Three Spaces of Optimization	p. 51
Fitness Landscapes	p. 52
High-Dimensional Cognitive Space and Word Meanings	p. 55
Two Factors of Complexity: NK Landscapes	p. 60
Combinatorial Optimization	p. 64
Binary Optimization	p. 67
Random and Greedy Searches	p. 71
Hill Climbing	p. 72
Simulated Annealing	p. 73
Binary and Gray Coding	p. 74
Step Sizes and Granularity	p. 75
Optimizing with Real Numbers	p. 77
Summary	p. 78
On Our Nonexistence as Entities: The Social Organism	p. 81
Views of Evolution	p. 82
Gaia: The Living Earth	p. 83
Differential Selection	p. 86
Our Microscopic Masters?	p. 91
Looking for the Right Zoom Angle	p. 92
Flocks, Herds, Schools, and Swarms: Social Behavior as Optimization	p. 94
Accomplishments of the Social Insects	p. 98
Optimizing with Simulated Ants: Computational Swarm Intelligence	p. 105
Staying Together but Not Colliding: Flocks, Herds, and Schools	p. 109
Robot Societies	p. 115
Shallow Understanding	p. 125
Agency	p. 129
Summary	p. 131
Evolutionary Computation Theory and Paradigms	p. 133
Introduction	p. 134
Evolutionary Computation History	p. 134
The Four Areas of Evolutionary Computation	p. 135
Genetic Algorithms	p. 135
Evolutionary Programming	p. 139
Evolution Strategies	p. 140
Genetic Programming	p. 141
Toward Unification	p. 141
Evolutionary Computation Overview	p. 142
EC Paradigm Attributes	p. 142
Implementation	p. 143
Genetic Algorithms	p. 146
An Overview	p. 146
A Simple GA Example Problem	p. 147
A Review of GA Operations	p. 152
Schemata and the Schema Theorem	p. 159
Final Comments on Genetic Algorithms	p. 163
Evolutionary Programming	p. 164
The Evolutionary Programming Procedure	p. 165
Finite State Machine Evolution	p. 166
Function Optimization	p. 169
Final Comments	p. 171
Evolution Strategies	p. 172
Mutation	p. 172
Recombination	p. 174
Selection	p. 175
Genetic Programming	p. 179
Summary	p. 185
Humans--Actual, Imagined, and Implied	p. 187
Studying Minds	p. 188
The Fall of the Behaviorist Empire	p. 193
The Cognitive Revolution	p. 195
Bandura's Social Learning Paradigm	p. 197
Social Psychology	p. 199
Lewin's Field Theory	p. 200
Norms, Conformity, and Social Influence	p. 202
Sociocognition	p. 205
Simulating Social Influence	p. 206
Paradigm Shifts in Cognitive Science	p. 210
The Evolution of Cooperation	p. 214
Explanatory Coherence	p. 216
Networks in Groups	p. 218
Culture in Theory and Practice	p. 220
Coordination Games	p. 223
The El Farol Problem	p. 226
Sugarscape	p. 229
Tesfatsion's ACE	p. 232
Picker's Competing-Norms Model	p. 233
Latane's Dynamic Social Impact Theory	p. 235
Boyd and Richerson's Evolutionary Culture Model	p. 240
Memetics	p. 245
Memetic Algorithms	p. 248
Cultural Algorithms	p. 253
Convergence of Basic and Applied Research	p. 254
Culture--and Life without It	p. 255
Summary	p. 258
Thinking Is Social	p. 261
Introduction	p. 262
Adaptation on Three Levels	p. 263
The Adaptive Culture Model	p. 263
Axelrod's Culture Model	p. 265
Similarity in Axelrod's Model	p. 267
Optimization of an Arbitrary Function	p. 268
A Slightly Harder and More Interesting Function	p. 269
A Hard Function	p. 271
Parallel Constraint Satisfaction	p. 273
Symbol Processing	p. 279
Discussion	p. 282
Summary	p. 284
The Particle Swarm and Collective Intelligence
The Particle Swarm	p. 287
Sociocognitive Underpinnings: Evaluate, Compare, and Imitate	p. 288
Evaluate	p. 288
Compare	p. 288
Imitate	p. 289
A Model of Binary Decision	p. 289
Testing the Binary Algorithm with the De Jong Test Suite	p. 297
No Free Lunch	p. 299
Multimodality	p. 302
Minds as Parallel Constraint Satisfaction Networks in Cultures	p. 307
The Particle Swarm in Continuous Numbers	p. 309
The Particle Swarm in Real-Number Space	p. 309
Pseudocode for Particle Swarm Optimization in Continuous Numbers	p. 313
Implementation Issues	p. 314
An Example: Particle Swarm Optimization of Neural Net Weights	p. 314
A Real-World Application	p. 318
The Hybrid Particle Swarm	p. 319
Science as Collaborative Search	p. 320
Emergent Culture, Immergent Intelligence	p. 323
Summary	p. 324
Variations and Comparisons	p. 327
Variations of the Particle Swarm Paradigm	p. 328
Parameter Selection	p. 328
Controlling the Explosion	p. 337
Particle Interactions	p. 342
Neighborhood Topology	p. 343
Substituting Cluster Centers for Previous Bests	p. 347
Adding Selection to Particle Swarms	p. 353
Comparing Inertia Weights and Constriction Factors	p. 354
Asymmetric Initialization	p. 357
Some Thoughts on Variations	p. 359
Are Particle Swarms Really a Kind of Evolutionary Algorithm?	p. 361
Evolution beyond Darwin	p. 362
Selection and Self-Organization	p. 363
Ergodicity: Where Can It Get from Here?	p. 366
Convergence of Evolutionary Computation and Particle Swarms	p. 367
Summary	p. 368
Applications	p. 369
Evolving Neural Networks with Particle Swarms	p. 370
Review of Previous Work	p. 370
Advantages and Disadvantages of Previous Approaches	p. 374
The Particle Swarm Optimization Implementation Used Here	p. 376
Implementing Neural Network Evolution	p. 377
An Example Application	p. 379
Conclusions	p. 381
Human Tremor Analysis	p. 382
Data Acquisition Using Actigraphy	p. 383
Data Preprocessing	p. 385
Analysis with Particle Swarm Optimization	p. 386
Summary	p. 389
Other Applications	p. 389
Computer Numerically Controlled Milling Optimization	p. 389
Ingredient Mix Optimization	p. 391
Reactive Power and Voltage Control	p. 391
Battery Pack State-of-Charge Estimation	p. 391
Summary	p. 392
Implications and Speculations	p. 393
Introduction	p. 394
Assertions	p. 395
Up from Social Learning: Bandura	p. 398
Information and Motivation	p. 399
Vicarious versus Direct Experience	p. 399
The Spread of Influence	p. 400
Machine Adaptation	p. 401
Learning or Adaptation?	p. 402
Cellular Automata	p. 403
Down from Culture	p. 405
Soft Computing	p. 408
Interaction within Small Groups: Group Polarization	p. 409
Informational and Normative Social Influence	p. 411
Self-Esteem	p. 412
Self-Attribution and Social Illusion	p. 414
Summary	p. 419
And in Conclusion...	p. 421
Statistics for Swarmers	p. 429
Genetic Algorithm Implementation	p. 451
Glossary	p. 457
References	p. 475
Index	p. 497
Table of Contents provided by Syndetics. All Rights Reserved.