Modern Heuristic Optimization Techniques

Kwang Y. Lee, PhD, is a Professor and Chair of Electrical and Computer Engineering at Baylor University (Texas). He is a Fellow of the IEEE. He was an associate editor of IEEE Transactions on Neural Networks and is an Editor of IEEE Transactions on Energy Conversion. He was also a member of the board of directors of the International Conference on Intelligent System Applications to Power Systems (ISAP).Mohamed A. El-Sharkawi, PhD, is a Professor of Electrical Engineering at the University of Washington. He is a Fellow of the IEEE, founder of the International Forum on the Application of Neural Networks to Power Systems (ANNPS), and cofounder of the International Conference on Intelligent System Applications to Power Systems (ISAP).

• Preface xxiContributors xxviiPart 1 Theory of Modern Heuristic Optimization 11 Introduction to Evolutionary Computation 3David B. Fogel1.1 Introduction 31.2 Advantages of Evolutionary Computation 41.2.1 Conceptual Simplicity 41.2.2 Broad Applicability 61.2.3 Outperform Classic Methods on Real Problems 71.2.4 Potential to Use Knowledge and Hybridize with Other Methods 81.2.5 Parallelism 81.2.6 Robust to Dynamic Changes 91.2.7 Capability for Self-Optimization 101.2.8 Able to Solve Problems That Have No Known Solutions 111.3 Current Developments 121.3.1 Review of Some Historical Theory in Evolutionary Computation 121.3.2 No Free Lunch Theorem 121.3.3 Computational Equivalence of Representations 141.3.4 Schema Theorem in the Presence of Random Variation 161.3.5 Two-Armed Bandits and the Optimal Allocation of Trials 171.4 Conclusions 19Acknowledgments 20References 202 Fundamentals of Genetic Algorithms 25Alexandre P. Alves da Silva and Djalma M. Falcao2.1 Introduction 252.2 Modern Heuristic Search Techniques 252.3 Introduction to GAs 272.4 Encoding 282.5 Fitness Function 302.5.1 Premature Convergence 322.5.2 Slow Finishing 322.6 Basic Operators 332.6.1 Selection 332.6.2 Crossover 362.6.3 Mutation 382.6.4 Control Parameters Estimation 382.7 Niching Methods 382.8 Parallel Genetic Algorithms 392.9 Final Comments 40Acknowledgments 41References 413 Fundamentals of Evolution Strategies and Evolutionary Programming 43Vladimiro Miranda3.1 Introduction 433.2 Evolution Strategies 463.2.1 The General (µ, κ, λ, ρ) Evolution Strategies Scheme 473.2.2 Some More Basic Concepts 503.2.3 The Early (1 + 1)ES and the 1/5 Rule 513.2.4 Focusing on the Optimum 533.2.5 The (1, λ)ES and σSA Self-Adaptation 543.2.6 How to Choose a Value for the Learning Parameter? 563.2.7 The (µ, l)ES as an Extension of (1, λ)ES 573.2.8 Self-Adaptation in (µ, λ)ES 583.3 Evolutionary Programming 603.3.1 The (µ + λ) Bridge to ES 603.3.2 A Scheme for Evolutionary Programming 613.3.3 Other Evolutionary Programming Variants 633.4 Common Features 633.4.1 Enhancing the Mutation Process 633.4.2 Recombination as a Major Factor 653.4.3 Handling Constraints 673.4.4 Starting Point 673.4.5 Fitness Function 673.4.6 Computing 683.5 Conclusions 68References 694 Fundamentals of Particle Swarm Optimization Techniques 71Yoshikazu Fukuyama4.1 Introduction 714.2 Basic Particle Swarm Optimization 724.2.1 Background of Particle Swarm Optimization 724.2.2 Original PSO 724.3 Variations of Particle Swarm Optimization 764.3.1 Discrete PSO 764.3.2 PSO for MINLPs 774.3.3 Constriction Factor Approach (CFA) 774.3.4 Hybrid PSO (HPSO) 784.3.5 Lbest Model 794.3.6 Adaptive PSO (APSO) 794.3.7 Evolutionary PSO (EPSO) 814.4 Research Areas and Applications 824.5 Conclusions 83References 835 Fundamentals of Ant Colony Search Algorithms 89Yong-Hua Song, Haiyan Lu, Kwang Y. Lee, and I. K. Yu5.1 Introduction 895.2 Ant Colony Search Algorithm 905.2.1 Behavior of Real Ants 905.2.2 Ant Colony Algorithms 915.2.3 Major Characteristics of Ant Colony Search Algorithms 985.3 Conclusions 99References 996 Fundamentals of Tabu Search 101Alcir J. Monticelli, Rubén Romero, and Eduardo Nobuhiro Asada6.1 Introduction 1016.1.1 Overview of the Tabu Search Approach 1016.1.2 Problem Formulation 1036.1.3 Coding and Representation 1046.1.4 Neighborhood Structure 1056.1.5 Characterization of the Neighborhood 1086.2 Functions and Strategies in Tabu Search 1106.2.1 Recency-Based Tabu Search 1106.2.2 Basic Tabu Search Algorithm 1126.2.3 The Use of Long-Term Memory in Tabu Search 1156.3 Applications of Tabu Search 1196.4 Conclusions 120References 1207 Fundamentals of Simulated Annealing 123Alcir J. Monticelli, Rubén Romero, and Eduardo Nobuhiro Asada7.1 Introduction 1237.2 Basic Principles 1257.2.1 Metropolis Algorithm 1257.2.2 Simulated Annealing Algorithm 1267.3 Cooling Schedule 1277.3.1 Determination of the Initial Temperature T0 1287.3.2 Determination of Nk 1297.3.3 Determination of Cooling Rate 1307.3.4 Stopping Criterion 1307.4 SA Algorithm for the Traveling Salesman Problem 1317.4.1 Problem Coding 1317.4.2 Evaluation of the Cost Function 1327.4.3 Cooling Schedule 1337.4.4 Comments on the Results for the TSP 1347.5 SA for Transmission Network Expansion Problem 1347.5.1 Problem Coding 1367.5.2 Determination of the Initial Solution 1367.5.3 Neighborhood Structure 1387.5.4 Variation of the Objective Function 1397.5.5 Cooling Schedule 1407.6 Parallel Simulated Annealing 1407.6.1 Division Algorithm 1417.6.2 Clustering Algorithm 1427.7 Applications of Simulated Annealing 1437.8 Conclusions 144References 1448 Fuzzy Systems 147Germano Lambert-Torres8.1 Motivation and Definitions 1478.1.1 Introduction 1478.1.2 Typical Actions in Fuzzy Systems 1488.2 Integration of Fuzzy Systems with Evolutionary Techniques 1508.2.1 Integration Types of Hybrid Systems 1508.2.2 Hybrid Systems in Evolutionary Techniques 1518.2.3 Evolutionary Algorithms and Fuzzy Logic 1528.3 An Illustrative Example of a Hybrid System 1528.3.1 Parking Conditions 1538.3.2 Creation of the Fuzzy Control 1548.3.3 First Simulations 1568.3.4 Problem Presentation 1568.3.5 Genetic Training Modulus Description 1588.3.6 The Option to Define the Starting Positions 1588.3.7 The Option Genetic Training 1588.3.8 Tests 1638.4 Conclusions 167References 1689 Differential Evolution, an Alternative Approach to Evolutionary Algorithm 171Kit Po Wong and ZhaoYang Dong9.1 Introduction 1719.2 Evolutionary Algorithms 1729.2.1 Basic EAs 1729.2.2 Virtual Population-Based Acceleration Techniques 1749.3 Differential Evolution 1769.3.1 Function Optimization Formulation 1769.3.2 DE Fundamentals 1779.4 Key Operators for Differential Evolution 1819.4.1 Encoding 1819.4.2 Mutation 1819.4.3 Crossover 1839.4.4 Other Operators 1839.5 An Optimization Example 1849.6 Conclusions 186Acknowledgments 186References 18610 Pareto Multiobjective Optimization 189Patrick N. Ngatchou, Anahita Zarei, Warren L. J. Fox, and Mohamed A. El-Sharkawi10.1 Introduction 18910.2 Basic Principles 19010.2.1 Generic Formulation of MO Problems 19110.2.2 Pareto Optimality Concepts 19110.2.3 Objectives of Multiobjective Optimization 19310.3 Solution Approaches 19410.3.1 Classic Methods 19410.3.2 Intelligent Methods 19610.4 Performance Analysis 20210.4.1 Objective of Performance Assessment 20210.4.2 Comparison Methodologies 20310.5 Conclusions 205Acknowledgments 205References 20511 Trust-Tech Paradigm for Computing High-Quality Optimal Solutions: Method and Theory 209Hsiao-Dong Chiang and Jaewook Lee11.1 Introduction 20911.2 Problem Preliminaries 21011.3 A Trust-Tech Paradigm 21311.3.1 Phase I 21311.3.2 Phase II 21411.4 Theoretical Analysis of Trust-Tech Method 21811.5 A Numerical Trust-Tech Method 22111.5.1 Computing Another Local Optimal Solution 22211.5.2 Computing Tier-One Local Optimal Solutions 22311.5.3 Computing Tier-N Solutions 22411.6 Hybrid Trust-Tech Methods 22511.7 Numerical Schemes 22711.8 Numerical Studies 22811.9 Conclusions Remarks 231References 232Part 2 Selected Applications of Modern Heuristic Optimization In Power Systems 23512 Overview of Applications in Power Systems 237Alexandre P. Alves da Silva, Djalma M. Falcão, and Kwang Y. Lee12.1 Introduction 23712.2 Optimization 23712.3 Power System Applications 23812.4 Model Identification 23912.4.1 Dynamic Load Modeling 23912.4.2 Short-Term Load Forecasting 24012.4.3 Neural Network Training 24112.5 Control 24212.5.1 Examples 24312.6 Distribution System Applications 24412.6.1 Network Reconfiguration for Loss Reduction 24512.6.2 Optimal Protection and Switching Devices Placement 24612.6.3 Prioritizing Investments in Distribution Networks 24712.7 Conclusions 249References 25013 Application of Evolutionary Technique to Power System Vulnerability Assessment 261Mingoo Kim, Mohamed A. El-Sharkawi, Robert J. Marks, and Ioannis N. Kassabalidis13.1 Introduction 26113.2 Vulnerability Assessment and Control 26313.3 Vulnerability Assessment Challenges 26413.3.1 Complexity of Power System 26413.3.2 VA On-line Speed 26513.3.3 Feature Selection 26513.3.4 Vulnerability Border 27013.3.5 Selection of Vulnerability Index 27613.4 Conclusions 281References 28114 Applications to System Planning 285Eduardo Nobuhiro Asada, Youngjae Jeon, Kwang Y. Lee, Vladimiro Miranda, Alcir J. Monticelli, Koichi Nara, Jong-Bae Park, Rubén Romero, and Yong-Hua Song14.1 Introduction 28514.2 Generation Expansion 28614.2.1 A Coding Strategy for an Improved GA for the Least-Cost GEP 28814.2.2 Fitness Function 28814.2.3 Creation of an Artificial Initial Population 28914.2.4 Stochastic Crossover Elitism and Mutation 29114.2.5 Numerical Examples 29214.2.6 Parameters for GEP and IGA 29314.2.7 Numerical Results 29514.3 Transmission Network Expansion 29714.3.1 Overview of Static Transmission Network Planning 29714.3.2 Solution Techniques for the Transmission Expansion Planning Problem 30014.3.3 Coding, Problem Representation, and Test Systems 30214.3.4 Complexity of the Test Systems 30414.3.5 Simulated Annealing 30614.3.6 Genetic Algorithms in Transmission Network Expansion Planning 30714.3.7 Tabu Search in Transmission Network Expansion Planning 30914.3.8 Hybrid TS/GA/SA Algorithm in Transmission Network Expansion Planning 31014.3.9 Comments on the Performance of Meta-heuristic Methods in Transmission Network Expansion Planning 31114.4 Distribution Network Expansion 31114.4.1 Dynamic Planning of Distribution System Expansion: A Complete GA Model 31214.4.2 Dynamic Planning of Distribution System Expansion: An Efficient GA Application 31614.4.3 Application of TS to the Design of Distribution Networks in FRIENDS 31714.5 Reactive Power Planning at Generation–Transmission Level 32014.5.1 Benders Decomposition of the Reactive Power Planning Problem 32114.5.2 Solution Algorithm 32314.5.3 Results for the IEEE 30-Bus System 32414.6 Reactive Power Planning at Distribution Level 32614.6.1 Modeling Chromosome Repair Using an Analytical Model 32614.6.2 Evolutionary Programming/Evolution Strategies Under Test 32714.7 Conclusions 330References 33015 Applications to Power System Scheduling 337Koay Chin Aik, Loi Lei Lai, Kwang Y. Lee, Haiyan Lu, Jong-Bae Park, Yong-Hua Song, Dipti Srinivasan, John G. Vlachogiannis, and I. K. Yu15.1 Introduction 33715.2 Economic Dispatch 33715.2.1 Economic Dispatch Problem 33715.2.2 GA Implementation to ED 33915.2.3 PSO Implementation to ED 34615.2.4 Numerical Example 34815.2.5 Summary 35415.3 Maintenance Scheduling 35415.3.1 Maintenance Scheduling Problem 35415.3.2 GA, PSO, and ES Implementation 35515.3.3 Simulation Results 36515.3.4 Summary 36615.4 Cogeneration Scheduling 36615.4.1 Cogeneration Scheduling Problem 36715.4.2 IGA Implementation 37015.4.3 Case Study 37315.4.4 Summary 37415.4.5 Nomenclature 37915.5 Short-Term Generation Scheduling of Thermal Units 38015.5.1 Short-Term Generation Scheduling Problem 38015.5.2 ACSA Implementation 38215.5.3 Experimental results 38515.6 Constrained Load Flow Problem 38515.6.1 Constrained Load Flow Problem 38515.6.2 Heuristic Ant Colony Search Algorithm Implementation 38615.6.3 Test Examples 39015.6.4 Summary 399References 39916 Power System Controls 403Yoshikazu Fukuyama, Hamid Ghezelayagh, Kwang Y. Lee, Chen-Ching Liu, Yong-Hua Song, and Ying Xiao16.1 Introduction 40316.2 Power System Controls: Particle Swarm Technique 40416.2.1 Problem Formulation of VVC 40516.2.2 Expansion of PSO for MINLP 40616.2.3 Voltage Security Assessment 40716.2.4 VVC Using PSO 40816.2.5 Numerical Examples 40916.2.6 Summary 41616.3 Power Plant Controller Design with GA 41716.3.1 Overview of the GA 41716.3.2 The Boiler-Turbine Model 41916.3.3 The GA Control System Design 42016.3.4 GA Design Results 42316.4 Evolutionary Programming Optimizer and Application in Intelligent Predictive Control 42716.4.1 Structure of the Intelligent Predictive Controller 42816.4.2 Power Plant Model 43016.4.3 Control Input Optimization 43116.4.4 Self-Organized Neuro-Fuzzy Identifier 43516.4.5 Rule Generation and Tuning 43816.4.6 Controller Implementation 44216.4.7 Summary 44416.5 An Interactive Compromise Programming-Based MO Approach to FACTS Control 44416.5.1 Review of MO Optimization Techniques 44616.5.2 Formulated MO Optimization Model 44916.5.3 Power Flow Control Model of FACTS Devices 45016.5.4 Proposed Interactive DWCP Method 45316.5.5 Proposed Interactive Procedure with Worst Compromise Displacement 45516.5.6 Implementation 45716.5.7 Numerical Results 45716.5.8 Summary 462References 46417 Genetic Algorithms for Solving Optimal Power Flow Problems 471Loi Lei Lai and Nidul Sinha17.1 Introduction 47117.2 Genetic Algorithms 47317.2.1 Terms Used in GA 47317.3 Load Flow Problem 47817.4 Optimal Power Flow Problem 48317.4.1 Application Examples 48517.5 OPF with FACTS Devices 48817.5.1 FACTS Model 49217.5.2 Problem Formulation 49517.5.3 Numerical Results 49617.6 Conclusions 499References 49918 An Interactive Compromise Programming-Based Multiobjective Approach to FACTS Control 501Ying Xiao, Yong-Hua Song, and Chen-Ching Liu18.1 Introduction 50118.2 Review of Multiobjective Optimization Techniques 50318.2.1 Weighting Method 50318.2.2 Goal Programming 50418.2.3 1-Constraint Method 50418.2.4 Compromise Programming 50418.2.5 Fuzzy Set Theory Applications 50518.2.6 Genetic Algorithm 50518.2.7 Interactive Procedure 50618.3 Formulated MO Optimization Model 50618.3.1 Formulated MO Optimization Model for FACTS Control 50718.3.2 Power Flow Control Model of FACTS Devices 50818.4 Proposed Interactive Displaced Worst Compromise Programming Method 51118.4.1 Applied Fuzzy CP 51118.4.2 Operation Cost Minimization 51218.4.3 Local Power Flow Control 51218.5 Proposed Interactive Procedure with WC Displacement 51318.5.1 Phase 1: Model Formulation 51318.5.2 Phase 2: Noninferior Solution Calculation 51418.5.3 Phase 3: Scenario Evaluation 51418.6 Implementation 51618.7 Numerical Results 51618.8 Conclusions 521References 52119 Hybrid Systems 525Vladimiro Miranda19.1 Introduction 52519.2 Capacitor Sizing and Location and Analytical Sensitivities 52719.2.1 From Darwin to Lamarck: Three Models 52819.2.2 Building a Lamarckian Acquisition of Improvements 52919.2.3 Analysis of a Didactic Example 53119.3 Unit Commitment Fuzzy Sets and Cleverer Chromosomes 53819.3.1 The Deceptive Characteristics of Unit Commitment Problems 53819.3.2 Similarity Between the Capacitor Placement and the Unit Commitment Problems 53919.3.3 The Need for Cleverer Chromosomes 54019.3.4 A Biological Touch: The Chromosome as a Program 54119.3.5 A Real-World Example: The CARE Model in Crete Greece 54219.3.6 Fitness Evaluation: Reliability (Spinning Reserve as a Fuzzy Constraint) 54719.3.7 Illustrative Results 54719.4 Voltage/Var Control and Loss Reduction in Distribution Networks with an Evolutionary Self-Adaptive Particle Swarm Optimization Algorithm: EPSO 55019.4.1 Justifying a Hybrid Approach 55019.4.2 The Principles of EPSO: Reproduction and Movement Rule 55119.4.3 Mutating Strategic Parameters 55219.4.4 The Merits of EPSO 55319.4.5 Experiencing with EPSO: Basic EPSO Model 55419.4.6 EPSO in Test Functions 55419.4.7 EPSO in Loss Reduction and Voltage/VAR Control: Definition of the Problem 55719.4.8 Applying EPSO in the Management of Networks with Distributed Generation 55819.5 Conclusions 559References 560Index 563

This text provides excellent, expert level, treatment of a very important systems engineering topic that will benefit students and practicing engineers. (IEEE Power Electronics Society Newsletter, 3rd Quarter, 2008)