Agent-Based Evolutionary Search (eBook)

Ruhul A. Sarker, Tapabrata Ray (Herausgeber)

eBook Download: PDF

2010
291 Seiten
Springer Berlin (Verlag)
978-3-642-13425-8 (ISBN)

Infos

Agent based evolutionary search is an emerging paradigm in computational int- ligence offering the potential to conceptualize and solve a variety of complex problems such as currency trading, production planning, disaster response m- agement, business process management etc. There has been a significant growth in the number of publications related to the development and applications of agent based systems in recent years which has prompted special issues of journals and dedicated sessions in premier conferences. The notion of an agent with its ability to sense, learn and act autonomously - lows the development of a plethora of efficient algorithms to deal with complex problems. This notion of an agent differs significantly from a restrictive definition of a solution in an evolutionary algorithm and opens up the possibility to model and capture emergent behavior of complex systems through a natural age- oriented decomposition of the problem space. While this flexibility of represen- tion offered by agent based systems is widely acknowledged, they need to be - signed for specific purposes capturing the right level of details and description. This edited volume is aimed to provide the readers with a brief background of agent based evolutionary search, recent developments and studies dealing with various levels of information abstraction and applications of agent based evo- tionary systems. There are 12 peer reviewed chapters in this book authored by d- tinguished researchers who have shared their experience and findings spanning across a wide range of applications.

Inhalt

Title 2
Preface 6
Contents 7
List of Contributors 9
Agent Based Evolutionary Approach: An Introduction 12
Introduction 12
Evolutionary Algorithms 14
Agent and Multi-Agent System 14
Integration of MAS and EAs 17
Agent-Based Evolutionary Algorithms 18
A Brief Description on the Content of This Book 19
References 20
Multi-Agent Evolutionary Model for Global Numerical Optimization 23
Introduction 24
Multi-Agent Genetic Algorithm 25
Multi-Agent Evolutionary Model for Decomposable Function Optimization 44
Hierarchy Multi-Agent Genetic Algorithm 48
Conclusions 56
References 57
An Agent Based Evolutionary Approach for Nonlinear Optimization with Equality Constraints 59
Introduction 59
Agent-Based Evolutionary Algorithms 62
Environment of Agents 62
Behavior of Agents 63
Learning of Agents 63
Reasoning Capability of Agents 63
Agent-Based Memetic Algorithms 64
Crossover 65
Life Span Learning Process 65
New Learning Process for Handling Equality Constraints 65
Pseudo Code of the Other LSLPs 68
The Algorithm 70
Constraint Handling 70
Experimental Studies 70
Initial Design Experience 70
Experimental Results and Discussion 71
Comparison with Other Algorithms 73
Effect of the New LSLP 76
Effect of Probability of Using LSLP 77
Conclusions 80
Appendix 81
$g$03 81
$g$05 81
$g$11 81
$g$13 82
B01 82
B02 82
References 83
Multiagent-Based Approach for Risk Analysis in Mission Capability Planning 87
Introduction 88
Background 89
Project Scheduling Problems 89
Overview 89
Resource Investment Problems 90
Mission Capability Planning 91
Overview of Capability Planning Process 91
Mission Capability Planning Problem 93
Mathematical Formulation of MCPP 94
A Multiagent-Based Framework 94
General Framework 95
Options Production Layer — OPL 95
Risk Tolerance Layer —RTL 96
Risk Simulation 96
Feedback from Agents to the Solutions 97
Case Study 98
Test Scenarios 98
Parameter Settings 98
Results and Discussion 99
The Effect of Feedback Mechanism: A Pilot Study 102
Conclusion 103
References 104
Agent Based Evolutionary Dynamic Optimization 107
Introduction 107
Proposed Agent Based Evolutionary Search Algorithm 108
The Framework of AES 108
Behaviors of Agents 111
Competitive Behavior 111
Statistics Based Learning Behavior 112
Two Diversity Maintaining Schemes on AES 113
Random Immigrants Method (RI) 114
Adaptive Dual Mapping Method(ADM) 114
The Dynamic Testing Suite 114
Stationary Test Problems 114
One-Max Function 115
Royal Road Function 115
Deceptive Function 116
Double Deceptive Function 116
Generating Dynamic Test Problems 116
Experimental Study 117
Experimental Setting 117
Experimental Results on DOPs 118
Conclusions 123
References 124
Divide and Conquer in Coevolution: A Difficult Balancing Act 127
Introduction 127
Background 129
Basic CCEA 129
Why Are CCEAs Attractive? 130
Shortcomings of Basic CCEA 132
Proposed CCEA with Adaptive Variable Partitioning Based on Correlation 134
Numerical Experiments 137
Results on 50D Test Problems 137
Results for 100D Problems 141
Variation in Performance of CCEA-AVP with Different Values of Correlation Threshold 144
Conclusions and Further Studies 145
References 147
Complex Emergent Behaviour from Evolutionary Spatial Animat Agents 149
Introduction 149
A Review of Animat Models 151
The Animat Model 153
Model Basics 153
Changes to the Model 156
Fine Tuning 156
Animat Model Observations 159
Evolution Algorithms 160
Evolution Experiments 161
Simulation 1 – The Control 162
Simulation 2 – Evolution by Crossover Only 162
Simulation 3 – Evolution by Crossover and Mutation 164
Simulation 4 – Evolution with Scarce Resorces 165
Discussion and Conclusions 166
References 168
An Agent-Based Parallel Ant Algorithm with an Adaptive Migration Controller 170
Introduction 170
A Brief Introduction to a Continuous Ant Algorithm 172
Initialization 172
Selection 172
Dump Operation and Pheromones Update 173
Random Search 173
Implementation of the Agent-Based Parallel Ant Algorithm (APAA) 174
Division of the Solution Vector 175
Stagnation-Based Asynchronous Migration Controller (SAMC) 176
Experiments and Discussions 178
Parameter Settings 179
Comparison of Solution Quality 180
Comparison of Convergence Speed 182
Conclusions 185
References 185
An Attempt to Stochastic Modeling of Memetic Systems 187
Motivation 187
EMAS Definition 190
EMAS Structure 190
EMAS State 191
EMAS Behavior 192
EMAS Actions 193
EMAS Dynamics 197
iEMAS Extension 198
iEMAS Structure 198
iEMAS State 199
iEMAS Behavior 199
iEMAS Actions 200
iEMAS Dynamics 205
Experimental Results 206
Conclusions 208
References 209
Searching for the Effective Bidding Strategy Using Parameter Tuning in Genetic Algorithm 211
Introduction 211
Genetic Algorithms 212
Related Work 214
Bidding Strategy Framework 215
Algorithm 220
Experimental Setting 221
Experimental Evaluation 224
Results and Discussion 224
Conclusion 233
References 233
$PSO$ (Particle Swarm Optimization): One Method, Many Possible Applications 237
$PSO$ and Evolutionary Search 237
$PSO$: Algorithms Inspired by Nature Twice 239
Definitions 240
$PSO$ and Multiobjective Optimization Problems 241
$PSO$: A Population-Based Technique 245
Case Studies 249
Looking for Resources 249
Microeconomy: General Equilibirum Theory 252
Tactical vs. Strategic Behavior 257
Conclusions 259
References 261
$VISPLORE$: Exploring Particle Swarms by Visual Inspection 263
Related Work 265
Particle Swarm Optimization 267
The $VISPLORE$ Toolkit 267
Visualization of a Particle 267
Visualization of a Population as a Collection of Particles 271
Visualization of an Experiment as a Collection of Populations 276
Visualization of Experiments as a Collection of Experiments 278
Searching in $VISPLORE$ 279
Different Views in $VISPLORE$ 281
Customizing Plots in $VISPLORE$ 282
$VISPLORE$ on the Foxholes Function 284
An Application Example: Soccer Kick Simulation 286
Conclusion 289
References 291
Index 1

Produkt-Details

• Erscheint lt. Verlag: 12.7.2010
• Reihe/Serie: Adaptation, Learning, and Optimization
• Reihe/Serie: Adaptation, Learning, and Optimization
• Zusatzinfo: 291 p. 48 illus. in color.
• Verlagsort: Berlin
• Sprache: englisch
• Themenwelt: Informatik ► Theorie / Studium ► Künstliche Intelligenz / Robotik
• Themenwelt: Mathematik / Informatik ► Mathematik
• Themenwelt: Technik
• Schlagworte: Agent-Based Evolutionary Search • Agents • algorithm • algorithms • Architecture • Calculus • Communication • Control • Evolution • evolutionary algorithm • evolutionary computation • Evolutionary Search • Genetic algorithms • learning • linear optimization • Model • Modeling • Optimization
• ISBN-10: 3-642-13425-4 / 3642134254
• ISBN-13: 978-3-642-13425-8 / 9783642134258