Computational Intelligence (eBook)
640 Seiten
Wiley (Verlag)
978-0-470-51250-0 (ISBN)
Engelbrecht provides readers with a wide knowledge of Computational Intelligence (CI) paradigms and algorithms; inviting readers to implement and problem solve real-world, complex problems within the CI development framework. This implementation framework will enable readers to tackle new problems without any difficulty through a single Java class as part of the CI library.
Key features of this second edition include:
- A tutorial, hands-on based presentation of the material.
- State-of-the-art coverage of the most recent developments in computational intelligence with more elaborate discussions on intelligence and artificial intelligence (AI).
- New discussion of Darwinian evolution versus Lamarckian evolution, also including swarm robotics, hybrid systems and artificial immune systems.
- A section on how to perform empirical studies; topics including statistical analysis of stochastic algorithms, and an open source library of CI algorithms.
- Tables, illustrations, graphs, examples, assignments, Java code implementing the algorithms, and a complete CI implementation and experimental framework.
Computational Intelligence: An Introduction, Second Edition is essential reading for third and fourth year undergraduate and postgraduate students studying CI. The first edition has been prescribed by a number of overseas universities and is thus a valuable teaching tool. In addition, it will also be a useful resource for researchers in Computational Intelligence and Artificial Intelligence, as well as engineers, statisticians, operational researchers, and bioinformaticians with an interest in applying AI or CI to solve problems in their domains.
Check out http://www.ci.cs.up.ac.za for examples, assignments and Java code implementing the algorithms.
Andries P. Engelbrecht is a full professor in Computer Science at the University of Pretoria, South Africa. He holds a PhD (Computer Science) from the University of Stellenbosch (1999) and has been actively involved in the research of computational intelligence since 1992. His group performs research in artificial neural networks, swarm intelligence, evolutionary computation, artificial immune systems, data and text mining, image analysis and multi-agent systems.? The research done is both theoretical where the objective is to develop new algorithms or to improve existing algorithms, and also application oriented, making use of techniques from computational intelligence to solve real-world problems. Professor Engelbrecht is actively involved in consultation to industry and contract research for industry.
Computational Intelligence: An Introduction, Second Edition offers an in-depth exploration into the adaptive mechanisms that enable intelligent behaviour in complex and changing environments. The main focus of this text is centred on the computational modelling of biological and natural intelligent systems, encompassing swarm intelligence, fuzzy systems, artificial neutral networks, artificial immune systems and evolutionary computation. Engelbrecht provides readers with a wide knowledge of Computational Intelligence (CI) paradigms and algorithms; inviting readers to implement and problem solve real-world, complex problems within the CI development framework. This implementation framework will enable readers to tackle new problems without any difficulty through a single Java class as part of the CI library. Key features of this second edition include: A tutorial, hands-on based presentation of the material. State-of-the-art coverage of the most recent developments in computational intelligence with more elaborate discussions on intelligence and artificial intelligence (AI). New discussion of Darwinian evolution versus Lamarckian evolution, also including swarm robotics, hybrid systems and artificial immune systems. A section on how to perform empirical studies; topics including statistical analysis of stochastic algorithms, and an open source library of CI algorithms. Tables, illustrations, graphs, examples, assignments, Java code implementing the algorithms, and a complete CI implementation and experimental framework. Computational Intelligence: An Introduction, Second Edition is essential reading for third and fourth year undergraduate and postgraduate students studying CI. The first edition has been prescribed by a number of overseas universities and is thus a valuable teaching tool. In addition, it will also be a useful resource for researchers in Computational Intelligence and Artificial Intelligence, as well as engineers, statisticians, operational researchers, and bioinformaticians with an interest in applying AI or CI to solve problems in their domains. Check out http://www.ci.cs.up.ac.za for examples, assignments and Java code implementing the algorithms.
Andries P. Engelbrecht is a full professor in Computer Science at the University of Pretoria, South Africa. He holds a PhD (Computer Science) from the University of Stellenbosch (1999) and has been actively involved in the research of computational intelligence since 1992. His group performs research in artificial neural networks, swarm intelligence, evolutionary computation, artificial immune systems, data and text mining, image analysis and multi-agent systems.? The research done is both theoretical where the objective is to develop new algorithms or to improve existing algorithms, and also application oriented, making use of techniques from computational intelligence to solve real-world problems. Professor Engelbrecht is actively involved in consultation to industry and contract research for industry.
Computational Intelligence 3
Contents 9
Figures 21
Tables 25
Algorithms 27
Preface 31
Part I INTRODUCTION 33
1 Introduction to Computational Intelligence 35
1.1 Computational Intelligence Paradigms 36
1.1.1 Artificial Neural Networks 37
1.1.2 Evolutionary Computation 40
1.1.3 Swarm Intelligence 41
1.1.4 Artificial Immune Systems 41
1.1.5 Fuzzy Systems 42
1.2 Short History 43
1.3 Assignments 45
Part II ARTIFICIAL NEURAL NETWORKS 47
2 The Artificial Neuron 49
2.1 Calculating the Net Input Signal 49
2.2 Activation Functions 50
2.3 Artificial Neuron Geometry 52
2.4 Artificial Neuron Learning 53
2.4.1 Augmented Vectors 55
2.4.2 Gradient Descent Learning Rule 56
2.4.3 Widrow-Hoff Learning Rule 57
2.4.4 Generalized Delta Learning Rule 57
2.4.5 Error-Correction Learning Rule 57
2.5 Assignments 57
3 Supervised Learning Neural Networks 59
3.1 Neural Network Types 59
3.1.1 Feedforward Neural Networks 60
3.1.2 Functional Link Neural Networks 61
3.1.3 Product Unit Neural Networks 62
3.1.4 Simple Recurrent Neural Networks 64
3.1.5 Time-Delay Neural Networks 66
3.1.6 Cascade Networks 67
3.2 Supervised Learning Rules 68
3.2.1 The Supervised Learning Problem 68
3.2.2 Gradient Descent Optimization 70
3.2.3 Scaled Conjugate Gradient 77
3.2.4 LeapFrog Optimization 81
3.2.5 Particle Swarm Optimization 81
3.3 Functioning of Hidden Units 81
3.4 Ensemble Neural Networks 83
3.5 Assignments 86
4 Unsupervised Learning Neural Networks 87
4.1 Background 87
4.2 Hebbian Learning Rule 88
4.3 Principal Component Learning Rule 90
4.4 Learning Vector Quantizer-I 91
4.5 Self-Organizing Feature Maps 94
4.5.1 Stochastic Training Rule 94
4.5.2 Batch Map 97
4.5.3 Growing SOM 97
4.5.4 Improving Convergence Speed 99
4.5.5 Clustering and Visualization 101
4.5.6 Using SOM 103
4.6 Assignments 103
5 Radial Basis Function Networks 105
5.1 Learning Vector Quantizer-II 105
5.2 Radial Basis Function Neural Networks 105
5.2.1 Radial Basis Function Network Architecture 106
5.2.2 Radial Basis Functions 107
5.2.3 Training Algorithms 108
5.2.4 Radial Basis Function Network Variations 112
5.3 Assignments 113
6 Reinforcement Learning 115
6.1 Learning through Awards 115
6.2 Model-Free Reinforcement Learning Model 118
6.2.1 Temporal Di.erence Learning 118
6.2.2 Q-Learning 118
6.3 Neural Networks and Reinforcement Learning 119
6.3.1 RPROP 119
6.3.2 Gradient Descent Reinforcement Learning 120
6.3.3 Connectionist Q-Learning 121
6.4 Assignments 123
7 Performance Issues (Supervised Learning) 125
7.1 Performance Measures 125
7.1.1 Accuracy 125
7.1.2 Complexity 130
7.1.3 Convergence 130
7.2 Analysis of Performance 130
7.3 Performance Factors 131
7.3.1 Data Preparation 131
7.3.2 Weight Initialization 138
7.3.3 Learning Rate and Momentum 139
7.3.4 Optimization Method 141
7.3.5 Architecture Selection 141
7.3.6 Adaptive Activation Functions 147
7.3.7 Active Learning 148
7.4 Assignments 156
Part III EVOLUTIONARY COMPUTATION 157
8 Introduction to Evolutionary Computation 159
8.1 Generic Evolutionary Algorithm 160
8.2 Representation – The Chromosome 161
8.3 Initial Population 164
8.4 Fitness Function 165
8.5 Selection 166
8.5.1 Selective Pressure 167
8.5.2 Random Selection 167
8.5.3 Proportional Selection 167
8.5.4 Tournament Selection 169
8.5.5 Rank-Based Selection 169
8.5.6 Boltzmann Selection 170
8.5.7 (µ +, ?)-Selection 171
8.5.8 Elitism 171
8.5.9 Hall of Fame 171
8.6 Reproduction Operators 171
8.7 Stopping Conditions 172
8.8 Evolutionary Computation versus Classical Optimization 173
8.9 Assignments 173
9 Genetic Algorithms 175
9.1 Canonical Genetic Algorithm 175
9.2 Crossover 176
9.2.1 Binary Representations 177
9.2.2 Floating-Point Representation 178
9.3 Mutation 185
9.3.1 Binary Representations 186
9.3.2 Floating-Point Representations 187
9.3.3 Macromutation Operator – Headless Chicken 188
9.4 Control Parameters 188
9.5 Genetic Algorithm Variants 189
9.5.1 Generation Gap Methods 190
9.5.2 Messy Genetic Algorithms 191
9.5.3 Interactive Evolution 193
9.5.4 Island Genetic Algorithms 194
9.6 Advanced Topics 196
9.6.1 Niching Genetic Algorithms 197
9.6.2 Constraint Handling 201
9.6.3 Multi-Objective Optimization 202
9.6.4 Dynamic Environments 205
9.7 Applications 206
9.8 Assignments 207
10 Genetic Programming 209
10.1 Tree-Based Representation 209
10.2 Initial Population 211
10.3 Fitness Function 212
10.4 Crossover Operators 212
10.5 Mutation Operators 214
10.6 Building Block Genetic Programming 216
10.7 Applications 216
10.8 Assignments 217
11 Evolutionary Programming 219
11.1 Basic Evolutionary Programming 219
11.2 Evolutionary Programming Operators 221
11.2.1 Mutation Operators 221
11.2.2 Selection Operators 225
11.3 Strategy Parameters 227
11.3.1 Static Strategy Parameters 227
11.3.2 Dynamic Strategies 227
11.3.3 Self-Adaptation 230
11.4 Evolutionary Programming Implementations 232
11.4.1 Classical Evolutionary Programming 232
11.4.2 Fast Evolutionary Programming 233
11.4.3 Exponential Evolutionary Programming 233
11.4.4 Accelerated Evolutionary Programming 233
11.4.5 Momentum Evolutionary Programming 234
11.4.6 Evolutionary Programming with Local Search 235
11.4.7 Evolutionary Programming with Extinction 235
11.4.8 Hybrid with Particle Swarm Optimization 236
11.5 Advanced Topics 238
11.5.1 Constraint Handling Approaches 238
11.5.2 Multi-Objective Optimization and Niching 238
11.5.3 Dynamic Environments 238
11.6 Applications 239
11.6.1 Finite-State Machines 239
11.6.2 Function Optimization 240
11.6.3 Training Neural Networks 241
11.6.4 Real-World Applications 242
11.7 Assignments 242
12 Evolution Strategies 245
12.1 (1+1)-ES 245
12.2 Generic Evolution Strategy Algorithm 247
12.3 Strategy Parameters and Self-Adaptation 248
12.3.1 Strategy Parameter Types 248
12.3.2 Strategy Parameter Variants 250
12.3.3 Self-Adaptation Strategies 251
12.4 Evolution Strategy Operators 253
12.4.1 Selection Operators 253
12.4.2 Crossover Operators 254
12.4.3 Mutation Operators 256
12.5 Evolution Strategy Variants 258
12.5.1 Polar Evolution Strategies 258
12.5.2 Evolution Strategies with Directed Variation 259
12.5.3 Incremental Evolution Strategies 260
12.5.4 Surrogate Evolution Strategy 261
12.6 Advanced Topics 261
12.6.1 Constraint Handling Approaches 261
12.6.2 Multi-Objective Optimization 262
12.6.3 Dynamic and Noisy Environments 265
12.6.4 Niching 265
12.7 Applications of Evolution Strategies 267
12.8 Assignments 267
13 Differential Evolution 269
13.1 Basic Differential Evolution 269
13.1.1 Difference Vectors 270
13.1.2 Mutation 271
13.1.3 Crossover 271
13.1.4 Selection 272
13.1.5 General Differential Evolution Algorithm 273
13.1.6 Control Parameters 273
13.1.7 Geometrical Illustration 274
13.2 DE/x/y/z 274
13.3 Variations to Basic Differential Evolution 277
13.3.1 Hybrid Differential Evolution Strategies 277
13.3.2 Population-Based Differential Evolution 282
13.3.3 Self-Adaptive Differential Evolution 282
13.4 Differential Evolution for Discrete-Valued Problems 284
13.4.1 Angle Modulated Differential Evolution 285
13.4.2 Binary Differential Evolution 286
13.5 Advanced Topics 287
13.5.1 Constraint Handling Approaches 288
13.5.2 Multi-Objective Optimization 288
13.5.3 Dynamic Environments 289
13.6 Applications 291
13.7 Assignments 291
14 Cultural Algorithms 293
14.1 Culture and Artificial Culture 293
14.2 Basic Cultural Algorithm 294
14.3 Belief Space 295
14.3.1 Knowledge Components 296
14.3.2 Acceptance Functions 297
14.3.3 Adjusting the Belief Space 298
14.3.4 Influence Functions 299
14.4 Fuzzy Cultural Algorithm 300
14.4.1 Fuzzy Acceptance Function 301
14.4.2 Fuzzified Belief Space 301
14.4.3 Fuzzy Influence Function 302
14.5 Advanced Topics 303
14.5.1 Constraint Handling 303
14.5.2 Multi-Objective Optimization 304
14.5.3 Dynamic Environments 305
14.6 Applications 306
14.7 Assignments 306
15 Coevolution 307
15.1 Coevolution Types 308
15.2 Competitive Coevolution 308
15.2.1 Competitive Fitness 309
15.2.2 Generic Competitive Coevolutionary Algorithm 311
15.2.3 Applications of Competitive Coevolution 312
15.3 Cooperative Coevolution 313
15.4 Assignments 315
Part IV COMPUTATIONAL SWARM INTELLIGENCE 317
16 Particle Swarm Optimization 321
16.1 Basic Particle Swarm Optimization 321
16.1.1 Global Best PSO 322
16.1.2 Local Best PSO 323
16.1.3 gbest versus lbest PSO 324
16.1.4 Velocity Components 325
16.1.5 Geometric Illustration 326
16.1.6 Algorithm Aspects 328
16.2 Social Network Structures 332
16.3 Basic Variations 335
16.3.1 Velocity Clamping 335
16.3.2 Inertia Weight 338
16.3.3 Constriction Coeffcient 341
16.3.4 Synchronous versus Asynchronous Updates 342
16.3.5 Velocity Models 342
16.4 Basic PSO Parameters 344
16.5 Single-Solution Particle Swarm Optimization 346
16.5.1 Guaranteed Convergence PSO 348
16.5.2 Social-Based Particle Swarm Optimization 349
16.5.3 Hybrid Algorithms 353
16.5.4 Sub-Swarm Based PSO 358
16.5.5 Multi-Start PSO Algorithms 365
16.5.6 Repelling Methods 369
16.5.7 Binary PSO 372
16.6 Advanced Topics 374
16.6.1 Constraint Handling Approaches 374
16.6.2 Multi-Objective Optimization 375
16.6.3 Dynamic Environments 378
16.6.4 Niching PSO 382
16.7 Applications 386
16.7.1 Neural Networks 386
16.7.2 Architecture Selection 388
16.7.3 Game Learning 388
16.8 Assignments 389
17 Ant Algorithms 391
17.1 Ant Colony Optimization Meta-Heuristic 392
17.1.1 Foraging Behavior of Ants 392
17.1.2 Stigmergy and Artificial Pheromone 395
17.1.3 Simple Ant Colony Optimization 396
17.1.4 Ant System 400
17.1.5 Ant Colony System 404
17.1.6 Max-Min Ant System 407
17.1.7 Ant-Q 410
17.1.8 Fast Ant System 411
17.1.9 Antabu 412
17.1.10 AS-rank 412
17.1.11 ANTS 413
17.1.12 Parameter Settings 415
17.2 Cemetery Organization and Brood Care 416
17.2.1 Basic Ant Colony Clustering Model 417
17.2.2 Generalized Ant Colony Clustering Model 418
17.2.3 Minimal Model for Ant Clustering 423
17.3 Division of Labor 423
17.3.1 Division of Labor in Insect Colonies 424
17.3.2 Task Allocation Based on Response Thresholds 425
17.3.3 Adaptive Task Allocation and Specialization 427
17.4 Advanced Topics 428
17.4.1 Continuous Ant Colony Optimization 428
17.4.2 Multi-Objective Optimization 430
17.4.3 Dynamic Environments 434
17.5 Applications 437
17.5.1 Traveling Salesman Problem 438
17.5.2 Quadratic Assignment Problem 439
17.5.3 Other Applications 443
17.6 Assignments 443
Part V ARTIFICIAL IMMUNE SYSTEMS 445
18 Natural Immune System 447
18.1 Classical View 447
18.2 Antibodies and Antigens 448
18.3 The White Cells 448
18.3.1 The Lymphocytes 449
18.4 Immunity Types 453
18.5 Learning the Antigen Structure 453
18.6 The Network Theory 454
18.7 The Danger Theory 454
18.8 Assignments 456
19 Artificial Immune Models 457
19.1 Artificial Immune System Algorithm 458
19.2 Classical View Models 460
19.2.1 Negative Selection 460
19.2.2 Evolutionary Approaches 461
19.3 Clonal Selection Theory Models 463
19.3.1 CLONALG 463
19.3.2 Dynamic Clonal Selection 465
19.3.3 Multi-Layered AIS 465
19.4 Network Theory Models 468
19.4.1 Artificial Immune Network 468
19.4.2 Self Stabilizing AIS 470
19.4.3 Enhanced Artificial Immune Network 472
19.4.4 Dynamic Weighted B-Cell AIS 473
19.4.5 Adapted Artificial Immune Network 474
19.4.6 aiNet 474
19.5 Danger Theory Models 477
19.5.1 Mobile Ad-Hoc Networks 477
19.5.2 An Adaptive Mailbox 478
19.5.3 Intrusion Detection 480
19.6 Applications and Other AIS models 480
19.7 Assignments 480
Part VI FUZZY SYSTEMS 483
20 Fuzzy Sets 485
20.1 Formal Definitions 486
20.2 Membership Functions 486
20.3 Fuzzy Operators 489
20.4 Fuzzy Set Characteristics 491
20.5 Fuzziness and Probability 494
20.6 Assignments 495
21 Fuzzy Logic and Reasoning 497
21.1 Fuzzy Logic 497
21.1.1 Linguistics Variables and Hedges 498
21.1.2 Fuzzy Rules 499
21.2 Fuzzy Inferencing 500
21.2.1 Fuzzification 501
21.2.2 Inferencing 502
21.2.3 Defuzzification 503
21.3 Assignments 504
22 Fuzzy Controllers 507
22.1 Components of Fuzzy Controllers 507
22.2 Fuzzy Controller Types 509
22.2.1 Table-Based Controller 509
22.2.2 Mamdani Fuzzy Controller 509
22.2.3 Takagi-Sugeno Controller 510
22.3 Assignments 510
23 Rough Sets 513
23.1 Concept of Discernibility 514
23.2 Vagueness in Rough Sets 515
23.3 Uncertainty in Rough Sets 516
23.4 Assignments 517
References 519
A Optimization Theory 583
A.1 Basic Ingredients of Optimization Problems 583
A.2 Optimization Problem Classifications 584
A.3 Optima Types 585
A.4 Optimization Method Classes 586
A.5 Unconstrained Optimization 587
A.5.1 Problem Definition 587
A.5.2 Optimization Algorithms 587
A.5.3 Example Benchmark Problems 591
A.6 Constrained Optimization 592
A.6.1 Problem Definition 592
A.6.2 Constraint Handling Methods 593
A.6.3 Example Benchmark Problems 598
A.7 Multi-Solution Problems 599
A.7.1 Problem Definition 599
A.7.2 Niching Algorithm Categories 600
A.7.3 Example Benchmark Problems 601
A.8 Multi-Objective Optimization 601
A.8.1 Multi-objective Problem 602
A.8.2 Weighted Aggregation Methods 603
A.8.3 Pareto-Optimality 604
A.9 Dynamic Optimization Problems 607
A.9.1 Definition 607
A.9.2 Dynamic Environment Types 608
A.9.3 Example Benchmark Problems 610
Index 613
| Erscheint lt. Verlag | 22.10.2007 |
|---|---|
| Sprache | englisch |
| Themenwelt | Informatik ► Theorie / Studium ► Künstliche Intelligenz / Robotik |
| Technik ► Elektrotechnik / Energietechnik | |
| Schlagworte | Computer Science • Electrical & Electronics Engineering • Elektrotechnik u. Elektronik • Fuzzy-Systeme • Fuzzy Systems • Informatik • Intelligente Systeme u. Agenten • Intelligent Systems & Agents • Künstliche Intelligenz • Künstliche Intelligenz • Neural networks • Neuronale Netze • Programmierung u. Software-Entwicklung • Programming & Software Development |
| ISBN-10 | 0-470-51250-4 / 0470512504 |
| ISBN-13 | 978-0-470-51250-0 / 9780470512500 |
| Haben Sie eine Frage zum Produkt? |
PDF (Adobe DRM)Größe: 4,8 MB
Kopierschutz: Adobe-DRM
Adobe-DRM ist ein Kopierschutz, der das eBook vor Mißbrauch schützen soll. Dabei wird das eBook bereits beim Download auf Ihre persönliche Adobe-ID autorisiert. Lesen können Sie das eBook dann nur auf den Geräten, welche ebenfalls auf Ihre Adobe-ID registriert sind.
Details zum Adobe-DRM
Dateiformat: PDF (Portable Document Format)
Mit einem festen Seitenlayout eignet sich die PDF besonders für Fachbücher mit Spalten, Tabellen und Abbildungen. Eine PDF kann auf fast allen Geräten angezeigt werden, ist aber für kleine Displays (Smartphone, eReader) nur eingeschränkt geeignet.
Systemvoraussetzungen:
PC/Mac: Mit einem PC oder Mac können Sie dieses eBook lesen. Sie benötigen eine
eReader: Dieses eBook kann mit (fast) allen eBook-Readern gelesen werden. Mit dem amazon-Kindle ist es aber nicht kompatibel.
Smartphone/Tablet: Egal ob Apple oder Android, dieses eBook können Sie lesen. Sie benötigen eine
Geräteliste und zusätzliche Hinweise
Zusätzliches Feature: Online Lesen
Dieses eBook können Sie zusätzlich zum Download auch online im Webbrowser lesen.
Buying eBooks from abroad
For tax law reasons we can sell eBooks just within Germany and Switzerland. Regrettably we cannot fulfill eBook-orders from other countries.
aus dem Bereich
