Evolutionary Computation: Principles and Practice for Signal Processing Book

	Preface	xi
Chapter 1	An Overview of Evolutionary Algorithms and Their Advantages	1
1.1	Introduction	1
1.2	Basics of Evolutionary Algorithms	1
1.3	Examples of Evolutionary Algorithm Applications	3
1.4	Advantages of Evolutionary Algorithms	9
1.4.1	Conceptual Simplicity	9
1.4.2	Broad Applicability	11
1.4.3	Outperform Classic Methods on Real Problems	11
1.4.4	Potential to Use Knowledge and Hybridize with Other Methods	12
1.4.5	Parallelism	13
1.4.6	Robust to Dynamic Changes	13
1.4.7	Capability for Self-Optimization	14
1.4.8	Able to Solve Problems with No Known Solutions	15
1.5	Principles and Practice of Evolutionary Algorithms	16
	References	16
Chapter 2	Evolving Models of Time Series	19
2.1	Introduction to Time Series Prediction and Modeling	19
2.2	Linear Models with Payoffs Other Than Least Squares	22
2.2.1	Examples of Evolving Polynomials for Arbitrary Criteria	23
2.3	Generating ARMA Models in One Step	27
2.3.1	The Traditional Two-Step Approach	27
2.3.2	Information Criteria for Model Building	29
2.3.3	Evolutionary Modeling	31
2.4	Neural Models	36
2.4.1	Evolving a Neural Predictor for a Chaotic Time Series	40
2.4.2	Evolving a Predictor of a Financial Time Series	42
2.5	Multiple Interactive Programs	48
2.6	Discussion	50
	References	54
Chapter 3	Evolutionary Clustering and Classification	57
3.1	Concepts of Clustering and Classification	57
3.2	Evolutionary Clustering	57
3.2.1	Method	59
3.2.2	Evaluating Solutions	62
3.2.3	Problems Investigated	63
3.2.4	Results and Discussion	63
3.3	Evolutionary Classification	69
3.3.1	Neural Networks for Classifying Sonar Data	69
3.3.1.1	Back Propagation	70
3.3.1.2	Simulated Annealing	71
3.3.1.3	Evolutionary Algorithms	72
3.3.1.4	Method and Materials	72
3.3.1.5	Experiments	74
3.3.1.6	Results	78
3.3.2	Neural Networks for Classifying Breast Cancer Data	82
3.3.2.1	Methods	82
3.3.2.1.1	Input Features and Neural Network Architecture	83
3.3.2.1.2	Training and Evaluation	84
3.3.2.2	Results	85
3.3.2.3	Discussion	88
3.4	Discussion	90
	References	90
Chapter 4	Evolving Control Systems	95
4.1	Introduction to Control	95
4.2	Evolutionary Control as Function Optimization	95
4.2.1	Three Textbook Problems	95
4.2.1.1	The Linear-Quadratic Problem	96
4.2.1.2	The Harvest Problem	97
4.2.1.3	The Push-Cart Problem	98
4.2.1.4	Experimental Results	98
4.2.2	Traffic Ramp Control	98
4.2.2.1	Background	98
4.2.2.2	Ramp-Metering Control Rules	105
4.2.2.3	Methods	108
4.2.2.4	Results	109
4.2.2.4.1	Scenario I: High Mainline and Ramp Demand	109
4.2.2.4.2	Scenario II: Moderate Mainline and Ramp Demand with an Incident	112
4.2.2.4.3	Scenario III: High Mainline and Ramp Demand with an Incident	116
4.2.2.5	Discussion	118
4.3	Evolutionary Control through Model Identification	118
4.3.1	Steps Toward Controlling Blood Pressure during Surgery	118
4.3.1.1	Introduction	118
4.3.1.2	Method, Materials, and Results	119
4.3.1.3	Discussion	124
4.3.2	Controlling a Pole Balanced on a Cart	128
4.3.2.1	System Dynamics	128
4.3.2.2	Method	129
4.3.2.3	Results	131
4.3.2.4	Discussion	134
4.4	Summary	134
	References	136
Chapter 5	Theory and Tools for Improving Evolutionary Algorithms	139
5.1	Theory	139
5.1.1	Binary Representations and Maximizing Implicit Parallelism	139
5.1.2	Crossover and Building Blocks	140
5.1.3	The Schema Theorem: The Fundamental Theorem of Genetic Algorithms	143
5.1.4	Proportional Selection and the K-Armed Bandit	144
5.1.5	A New Direction	145
5.2	Methods to Relate Parent and Offspring Fitness	146
5.3	Experiments with Fitness Distributions	148
5.3.1	Methods	148
5.3.2	Results	152
5.4	Discussion	157
	References	159
	Index	163