Neural Networks and Learning Algorithms in MATLAB
Neural Networks and Learning Algorithms in MATLAB
Ardahir Mohammadazadeh, Mohammad Hosein Sabzalian, Oscar Castillo, Rathinasamy Sakthivel, Fayez F. M. El-Sousy, Saleh Mobayen
Contents
1 Introduction . 1
1.1 Overview . 1
1.2 Some Applications of Neural Networks . 2
1.3 Different Types of Neural Network Training . 3
1.4 Learning Principles in Neural Networks . 4
References . 4
2 Multilayer Perceptron (MLP) Neural Networks . 5
2.1 Training Based on Error Backpropagation . 8
2.2 Implementation in MATLAB 9
2.3 Application of Neural Network in Classification 12
2.4 Over-Parameterization 15
2.5 Over-Training 18
2.6 Training Based on Full Propagation 18
References . 21
3 Neural Networks Training Based on Recursive Least Squares (RLS) 23
3.1 RLS Training Technique 23
3.2 Implementation in MATLAB 24
3.3 Comparison with Gradient Descent . 29
4 Neural Networks Training Based on Second-Order Optimization
Technique . 31
4.1 Introduction . 31
4.2 Newton’s Method 32
4.3 Levenberg–Marquardt Algorithm . 32
4.4 Conjugate Gradient (CG) Method 33
4.5 Implementation in MATLAB 33
References . 38
viiviii Contents
5 Neural Networks Training Based on Genetic Algorithm 39
5.1 Introduction . 39
5.1.1 What is the Genetic Algorithm (GA)? . 39
5.1.2 Operators of a Genetic Algorithm . 41
5.1.3 Applications of Genetic Algorithm 42
5.2 Genetic Algorithm in MATLAB 43
5.3 Optimization of Neural Network Parameters Based on Genetic
Algorithm . 54
Reference 59
6 Neural Network Training Based Particle Swarm Optimization (PSO) . 61
6.1 Introduction . 61
6.2 Algorithm Formulation 61
6.3 Implementation in MATLAB 64
References . 68
7 Neural Network Training Based on UKF . 69
7.1 UKF Algorithm 69
7.2 Implementation in MATLAB 72
References . 78
8 Designing Neural-Fuzzy PID Controller Through Multiobjective
Optimization . 79
8.1 Introduction . 79
8.2 Classic Methods . 79
8.2.1 Ziegler–Nichols Method . 79
8.2.2 Cohen-Coon Method 80
8.2.3 Smart Methods . 80
8.2.4 Single-Objective Optimization 81
8.2.5 Multiobjective Optimization 82
8.2.6 Primary Definitions . 83
8.2.7 Decision Variables 84
8.2.8 Constraints . 84
8.2.9 Objective Functions . 84
8.2.10 Dominance 84
8.2.11 Non-Dominated Set . 85
8.2.12 Pareto Principle 85
8.2.13 Optimal Pareto Solution . 86
8.2.14 Optimal Pareto Set 86
8.3 Objectives of Multiobjective Optimization . 87
8.3.1 Common Algorithms in Solving Multiobjective
Optimization . 87
8.4 Designing Multiobjective PID Controller 88Contents ix
8.5 Designing a MOPID Controller for a Sample Power System . 89
8.5.1 First State . 90
8.5.2 Second State . 93
8.6 Using Fuzzy-Neural Network for Gain Schedule . 94
8.7 Fuzzy-Neural Network Training for PID Controller Regulation . 96
8.7.1 Simulation for Fuzzy-Neural Controller of Gain Schedule 98
8.8 Conclusion 100
8.9 Implementation in MATLAB 101
8.9.1 Dynamic Model of Power System 101
8.9.2 First Example 105
8.9.3 Supplementary Ideas on Modeling the Power System
for the Frequency Load Problem 106
Uncited Reference
كلمة سر فك الضغط : books-world.net
The Unzip Password : books-world.net
تعليقات