书籍详情
极值优化：原理、算法和应用（英文版）
作者：吕勇哉，陈玉旺，陈泯融，陈鹏，曾国强 著
出版社：化学工业出版社
出版时间：2016-08-01
ISBN：9787122270016
定价：¥128.00
购买这本书可以去
内容简介
本书在总结编者近年来的原创性成果的基础上，综合了大量的EO相关文献，从原理、算法和应用等方面来介绍EO算法，内容横跨了多个学科，如运筹学、计算机软件、系统控制和制造工业等。本书的内容主要包括了以下四个方面。（1）本书重点探讨了具有较高计算复杂度的优化问题的解决方法，并对这些优化方法进行了归纳和总结。（2）针对一些标准测试问题，本书从原理、工作机理、算法和仿真实验等方面对EO算法内在的极值动力学机制及其应用进行了全面的介绍。另外，本书将EO算法与一些经典的启发式算法进行了仿真比较。（3）在总结编者近年来对EO的原创性研究成果的基础上，本书重点介绍了EO算法在自组织优化、进化概率分布和结构特征（例如骨架）等方面的工作机理。本书还介绍了各种改进的EO算法和基于EO的混合计算智能方法。（4）本书将EO算法和改进的EO算法应用于实际的工程领域，例如，多目标优化领域、生物信息学领域、系统建模和控制领域以及生产调度领域。本书对于从事自动控制优化工作的研究人员以及工程技术人员学习和掌握EO方法具有重要作用。
作者简介
吕勇哉，浙江大学，教授，吕勇哉教授，美国电工电子工程师学会（IEEE） Fellow、国际自动控制联合会（IFAC）主席 （1996年~1999年）。成功地领导和研究开发了集成神经网络控制系统和遗传算法优化调度等系统。曾获美国仪器仪表学会UOP技术奖，于1995年和1996年连获美国钢铁学会Kelly奖，出版了《Industrial Intelligent Control： Fundamentals and Applications》等专著，并获得国家科学技术进步二等奖，全国科技图书一等奖，国家教委教学成果奖和多项部级奖。曾任中国自动化学会副理事长、国务院学位委员会和国家自然科学基金会自动化学科评审组成员和浙江大学学术委员会副主任等职。
目录
Prefacexi
Acknowledgmentsxv
Section I FUNDAMENT ALS, MET HODOLOGY,AND ALGORIT HMS
1 General Introduction3
1.1 Introduction3
1.2 Understanding Optimization: From Practical Aspects4
1.2.1 Mathematical Optimization4
1.2.2 Optimization: From Practical Aspects5
1.2.3 Example Applications of Optimization6
1.2.4 Problem Solving for Optimization8
1.3 Phase Transition and Computational Complexity9
1.3.1 Computational Complexity in General9
1.3.2 Phase Transition in Computation10
1.4 CI-Inspired Optimization11
1.4.1 Evolutionary Computations11
1.4.2 Swarm Intelligence12
1.4.3 Data Mining and Machine Learning13
1.4.4 Statistical Physics13
1.5 Highlights of EO14
1.5.1 Self-Organized Criticality and EO14
1.5.2 Coevolution, Ecosystems, and Bak–Sneppen Model16
1.5.3 Comparing EO with SA and GA17
1.5.4 Challenging Open Problems17
1.6 Organization of the Book18
2 Introduction to Extremal Optimization21
2.1 Optimization with Extremal Dynamics21
2.2 Multidisciplinary Analysis of EO23
2.3 Experimental and Comparative Analysis on the Traveling Salesman Problems24
2.3.1 EO for the Symmetric TSP25
2.3.1.1 Problem Formulation and Algorithm Design25
2.3.2 SA versus Extremal Dynamics27
2.3.3 Optimizing Near the Phase Transition30
2.3.4 EO for the Asymmetric TSP31
2.3.4.1 Cooperative Optimization32
2.3.4.2 Parameter Analysis33
2.4 Summary35
3 Extremal Dynamics–Inspired Self-Organizing Optimization37
3.1 Introduction37
3.2 Analytic Characterization of COPs39
3.2.1 Modeling COPs into Multientity Systems39
3.2.2 Local Fitness Function 40
3.2.3 Microscopic Analysis of Optimal Solutions43
3.2.4 Neighborhood and Fitness Network 46
3.2.5 Computational Complexity and Phase Transition49
3.3 Self-Organized Optimization51
3.3.1 Self-Organized Optimization Algorithm51
3.3.2 Comparison with Related Methods53
3.3.2.1 Simulated Annealing54
3.3.2.2 Genetic Algorithm54
3.3.2.3 Extremal Optimization55
3.3.3 Experimental Validation55
3.4 Summary57
Section II MODIFIED EO AND INTE GRATION OF EO WITH OTHER SOLUTIONS TO COMPUTATIONAL INTELLIGENCE
4 Modified Extremal Optimization61
4.1 Introduction61
4.2 Modified EO with Extended Evolutionary Probability Distribution61
4.2.1 Evolutionary Probability Distribution62
4.2.2 Modified EO Algorithm with Extended Evolutionary Probability Distribution 64
4.2.3 Experimental Results67
4.3 Multistage EO70
4.3.1 Motivations70
4.3.2 MSEO Algorithm72
4.3.3 Experimental Results73
4.3.3.1 The Simplest Case: Two-Stage EO73
4.3.3.2 Complex Case74
4.3.4 Adjustable Parameters versus Performance 77
4.4 Backbone-Guided EO78
4.4.1 Definitions of Fitness and Backbones80
4.4.2 BGEO Algorithm81
4.4.3 Experimental Results84
4.5 Population-Based EO88
4.5.1 Problem Formulation of Numerical Constrained Optimization Problems90
4.5.2 PEO Algorithm91
4.5.3 Mutation Operator92
4.5.4 Experimental Results94
4.5.5 Advantages of PEO97
4.6 Summary98
5 Memetic Algorithms with Extremal Optimization101
5.1 Introduction to MAs101
5.2 Design Principle of MAs102
5.3 EO–LM Integration105
5.3.1 Introduction105
5.3.2 Problem Statement and Math Formulation107
5.3.3 Introduction of LM GS108
5.3.4 MA-Based Hybrid EO–LM Algorithm109
5.3.5 Fitness Function113
5.3.6 Experimental Tests on Benchmark Problems 114
5.3.6.1 A Multi-Input, Single-Output Static Nonlinear Function 114
5.3.6.2 Five-Dimensional Ackley Function Regression 116
5.3.6.3 Dynamic Modeling for Continuously Stirred Tank Reactor 116
5.4 EO–SQP Integration119
5.4.1 Introduction 119
5.4.2 Problem Formulation121
5.4.3 Introduction of SQP122
5.4.4 MA-Based Hybrid EO–SQP Algorithm123
5.4.5 Fitness Function Definition125
5.4.6 Termination Criteria125
5.4.7 Workflow and Algorithm126
5.4.8 Experimental Tests on Benchmark Functions127
5.4.8.1 Unconstrained Problems128
5.4.8.2 Constrained Problems132
5.4.9 Dynamics Analysis of the Hybrid EO–SQP136
5.5 EO–PSO Integration138
5.5.1 Introduction138
5.5.2 Particle Swarm Optimization139
5.5.3 PSO–EO Algorithm140
5.5.4 Mutation Operator140
5.5.5 Computational Complexity143
5.5.6 Experimental Results143
5.6 EO–ABC Integration149
5.6.1 Artificial Bee Colony150
5.6.2 ABC–EO Algorithm153
5.6.3 Mutation Operator154
5.6.4 Differences between ABC–EO and Other Hybrid Algorithms155
5.6.5 Experimental Results 155
5.7 EO–GA Integration160
5.8 Summary163
6 Multiobjective Optimization with Extremal Dynamics165
6.1 Introduction165
6.2 Problem Statement and Definition167
6.3 Solutions to Multiobjective Optimization168
6.3.1 Aggregating Functions168
6.3.2 Population-Based Non-Pareto Approaches169
6.3.3 Pareto-Based Approaches169
6.4 EO for Numerical MOPs170
6.4.1 MOEO Algorithm171
6.4.1.1 Fitness Assignment171
6.4.1.2 Diversity Preservation173
6.4.1.3 External Archive174
6.4.1.4 Mutation Operation175
6.4.2 Unconstrained Numerical MOPs with MOEO176
6.4.2.1 Performance Metrics176
6.4.2.2 Experimental Settings179
6.4.2.3 Experimental Results and Discussion179
6.4.2.4 Conclusions185
6.4.3 Constrained Numerical MOPs with MOEO185
6.4.3.1 Performance Metrics186
6.4.3.2 Experimental Settings187
6.4.3.3 Experimental Results and Discussion188
6.4.3.4 Conclusions 191
6.5 Multiobjective 0/1 Knapsack Problem with MOEO 191
6.5.1 Extended MOEO for MOKP 191
6.5.1.1 Mutation Operation 191
6.5.1.2 Repair Strategy192
6.5.2 Experimental Settings193
6.5.3 Experimental Results and Discussion194
6.5.4 Conclusions195
6.6 Mechanical Components Design with MOEO197
6.6.1 Introduction197
6.6.2 Experimental Settings198
6.6.2.1 Two-Bar Truss Design (Two Bar for Short)198
6.6.2.2 Welded Beam Design (Welded Beam for Short).198
6.6.2.3 Machine Tool Spindle Design (Spindle for Short)199
6.6.3 Experimental Results and Discussion201
6.6.4 Conclusions202
6.7 Portfolio Optimization with MOEO203
6.7.1 Portfolio Optimization Model203
6.7.2 MOEO for Portfolio Optimization Problems205
6.7.2.1 Mutation Operation206
6.7.2.2 Repair Strategy207
6.7.3 Experimental Settings207
6.7.4 Experimental Results and Discussion 208
6.7.5 Conclusions212
6.8 Summary212
Section III APPLICATION S
7 EO for Systems Modeling and Control215
7.1 Problem Statement215
7.2 Endpoint Quality Prediction of Batch Production with MA-EO 216
7.3 EO for Kernel Function and Parameter Optimization in Support Vector Regression 219
7.3.1 Introduction221
7.3.2 Problem Formulation221
7.3.2.1 Support Vector Regression 222
7.3.2.2 Optimization of SVR Kernel Function and Parameters223
7.3.3 Hybrid EO-Based Optimization for SVR Kernel Function and Parameters224
7.3.3.1 Chromosome Structure224
7.3.3.2 Fitness Function225
7.3.3.3 EO-SVR Workflow 226
7.3.4 Experimental Results 228
7.3.4.1 Approximation of Single-Variable Function228
7.3.4.2 Approximation of Multivariable Function233
7.4 Nonlinear Model Predictive Control with MA-EO238
7.4.1 Problem Formulation for NMPC Based on SVM Model239
7.4.2 Real-Time NMPC with SVM and EO-SQP242
7.4.2.1 Workflow of Proposed NMPC242
7.4.2.2 Encoding Strategy244
7.4.2.3 Selection of the Initial Population246
7.4.2.4 Termination Criteria of the NLP Solver246
7.4.2.5 Horizon-Based EO Mutation for NMPC Online Optimization 246
7.4.3 Simulation Studies247
7.5 Intelligent PID Control with Binary-Coded EO252
7.5.1 PID Controllers and Performance Indices252
7.5.2 BCEO Algorithm255
7.5.3 Experimental Results258
7.5.3.1 Single-Variable Controlled Plant258
7.5.3.2 Multivariable Controlled Plant261
7.5.3.3 Parameters and Control Performances267
7.6 Summary267
8 EO for Production Planning and Scheduling271
8.1 Introduction271
8.1.1 An Overview of HSM Scheduling272
8.1.2 Production Process273
8.1.3 Scheduling Objectives and Constraints274
8.2 Problem Formulation276
8.3 Hybrid Evolutionary Solutions with the Integration of GA and EO 280
8.3.1 Global Search Algorithm: Modified GA for Order Selection and Sequencing 280
8.3.1.1 Representation of Solutions 280
8.3.1.2 Population Initialization281
8.3.1.3 Fitness Function282
8.3.1.4 Genetic Operators282
8.3.2 Local Improving Algorithm: τ-EO285
8.3.2.1 Introduction to EO 286
8.3.2.2 EO for Improving the Body Section286
8.3.2.3 Hybrid Evolutionary Algorithms288
8.3.3 Design of a HSM-Scheduling System291
8.3.4 Computational Results293
8.4 Summary296
References297
Author Index315
Subject Index323
Acknowledgmentsxv
Section I FUNDAMENT ALS, MET HODOLOGY,AND ALGORIT HMS
1 General Introduction3
1.1 Introduction3
1.2 Understanding Optimization: From Practical Aspects4
1.2.1 Mathematical Optimization4
1.2.2 Optimization: From Practical Aspects5
1.2.3 Example Applications of Optimization6
1.2.4 Problem Solving for Optimization8
1.3 Phase Transition and Computational Complexity9
1.3.1 Computational Complexity in General9
1.3.2 Phase Transition in Computation10
1.4 CI-Inspired Optimization11
1.4.1 Evolutionary Computations11
1.4.2 Swarm Intelligence12
1.4.3 Data Mining and Machine Learning13
1.4.4 Statistical Physics13
1.5 Highlights of EO14
1.5.1 Self-Organized Criticality and EO14
1.5.2 Coevolution, Ecosystems, and Bak–Sneppen Model16
1.5.3 Comparing EO with SA and GA17
1.5.4 Challenging Open Problems17
1.6 Organization of the Book18
2 Introduction to Extremal Optimization21
2.1 Optimization with Extremal Dynamics21
2.2 Multidisciplinary Analysis of EO23
2.3 Experimental and Comparative Analysis on the Traveling Salesman Problems24
2.3.1 EO for the Symmetric TSP25
2.3.1.1 Problem Formulation and Algorithm Design25
2.3.2 SA versus Extremal Dynamics27
2.3.3 Optimizing Near the Phase Transition30
2.3.4 EO for the Asymmetric TSP31
2.3.4.1 Cooperative Optimization32
2.3.4.2 Parameter Analysis33
2.4 Summary35
3 Extremal Dynamics–Inspired Self-Organizing Optimization37
3.1 Introduction37
3.2 Analytic Characterization of COPs39
3.2.1 Modeling COPs into Multientity Systems39
3.2.2 Local Fitness Function 40
3.2.3 Microscopic Analysis of Optimal Solutions43
3.2.4 Neighborhood and Fitness Network 46
3.2.5 Computational Complexity and Phase Transition49
3.3 Self-Organized Optimization51
3.3.1 Self-Organized Optimization Algorithm51
3.3.2 Comparison with Related Methods53
3.3.2.1 Simulated Annealing54
3.3.2.2 Genetic Algorithm54
3.3.2.3 Extremal Optimization55
3.3.3 Experimental Validation55
3.4 Summary57
Section II MODIFIED EO AND INTE GRATION OF EO WITH OTHER SOLUTIONS TO COMPUTATIONAL INTELLIGENCE
4 Modified Extremal Optimization61
4.1 Introduction61
4.2 Modified EO with Extended Evolutionary Probability Distribution61
4.2.1 Evolutionary Probability Distribution62
4.2.2 Modified EO Algorithm with Extended Evolutionary Probability Distribution 64
4.2.3 Experimental Results67
4.3 Multistage EO70
4.3.1 Motivations70
4.3.2 MSEO Algorithm72
4.3.3 Experimental Results73
4.3.3.1 The Simplest Case: Two-Stage EO73
4.3.3.2 Complex Case74
4.3.4 Adjustable Parameters versus Performance 77
4.4 Backbone-Guided EO78
4.4.1 Definitions of Fitness and Backbones80
4.4.2 BGEO Algorithm81
4.4.3 Experimental Results84
4.5 Population-Based EO88
4.5.1 Problem Formulation of Numerical Constrained Optimization Problems90
4.5.2 PEO Algorithm91
4.5.3 Mutation Operator92
4.5.4 Experimental Results94
4.5.5 Advantages of PEO97
4.6 Summary98
5 Memetic Algorithms with Extremal Optimization101
5.1 Introduction to MAs101
5.2 Design Principle of MAs102
5.3 EO–LM Integration105
5.3.1 Introduction105
5.3.2 Problem Statement and Math Formulation107
5.3.3 Introduction of LM GS108
5.3.4 MA-Based Hybrid EO–LM Algorithm109
5.3.5 Fitness Function113
5.3.6 Experimental Tests on Benchmark Problems 114
5.3.6.1 A Multi-Input, Single-Output Static Nonlinear Function 114
5.3.6.2 Five-Dimensional Ackley Function Regression 116
5.3.6.3 Dynamic Modeling for Continuously Stirred Tank Reactor 116
5.4 EO–SQP Integration119
5.4.1 Introduction 119
5.4.2 Problem Formulation121
5.4.3 Introduction of SQP122
5.4.4 MA-Based Hybrid EO–SQP Algorithm123
5.4.5 Fitness Function Definition125
5.4.6 Termination Criteria125
5.4.7 Workflow and Algorithm126
5.4.8 Experimental Tests on Benchmark Functions127
5.4.8.1 Unconstrained Problems128
5.4.8.2 Constrained Problems132
5.4.9 Dynamics Analysis of the Hybrid EO–SQP136
5.5 EO–PSO Integration138
5.5.1 Introduction138
5.5.2 Particle Swarm Optimization139
5.5.3 PSO–EO Algorithm140
5.5.4 Mutation Operator140
5.5.5 Computational Complexity143
5.5.6 Experimental Results143
5.6 EO–ABC Integration149
5.6.1 Artificial Bee Colony150
5.6.2 ABC–EO Algorithm153
5.6.3 Mutation Operator154
5.6.4 Differences between ABC–EO and Other Hybrid Algorithms155
5.6.5 Experimental Results 155
5.7 EO–GA Integration160
5.8 Summary163
6 Multiobjective Optimization with Extremal Dynamics165
6.1 Introduction165
6.2 Problem Statement and Definition167
6.3 Solutions to Multiobjective Optimization168
6.3.1 Aggregating Functions168
6.3.2 Population-Based Non-Pareto Approaches169
6.3.3 Pareto-Based Approaches169
6.4 EO for Numerical MOPs170
6.4.1 MOEO Algorithm171
6.4.1.1 Fitness Assignment171
6.4.1.2 Diversity Preservation173
6.4.1.3 External Archive174
6.4.1.4 Mutation Operation175
6.4.2 Unconstrained Numerical MOPs with MOEO176
6.4.2.1 Performance Metrics176
6.4.2.2 Experimental Settings179
6.4.2.3 Experimental Results and Discussion179
6.4.2.4 Conclusions185
6.4.3 Constrained Numerical MOPs with MOEO185
6.4.3.1 Performance Metrics186
6.4.3.2 Experimental Settings187
6.4.3.3 Experimental Results and Discussion188
6.4.3.4 Conclusions 191
6.5 Multiobjective 0/1 Knapsack Problem with MOEO 191
6.5.1 Extended MOEO for MOKP 191
6.5.1.1 Mutation Operation 191
6.5.1.2 Repair Strategy192
6.5.2 Experimental Settings193
6.5.3 Experimental Results and Discussion194
6.5.4 Conclusions195
6.6 Mechanical Components Design with MOEO197
6.6.1 Introduction197
6.6.2 Experimental Settings198
6.6.2.1 Two-Bar Truss Design (Two Bar for Short)198
6.6.2.2 Welded Beam Design (Welded Beam for Short).198
6.6.2.3 Machine Tool Spindle Design (Spindle for Short)199
6.6.3 Experimental Results and Discussion201
6.6.4 Conclusions202
6.7 Portfolio Optimization with MOEO203
6.7.1 Portfolio Optimization Model203
6.7.2 MOEO for Portfolio Optimization Problems205
6.7.2.1 Mutation Operation206
6.7.2.2 Repair Strategy207
6.7.3 Experimental Settings207
6.7.4 Experimental Results and Discussion 208
6.7.5 Conclusions212
6.8 Summary212
Section III APPLICATION S
7 EO for Systems Modeling and Control215
7.1 Problem Statement215
7.2 Endpoint Quality Prediction of Batch Production with MA-EO 216
7.3 EO for Kernel Function and Parameter Optimization in Support Vector Regression 219
7.3.1 Introduction221
7.3.2 Problem Formulation221
7.3.2.1 Support Vector Regression 222
7.3.2.2 Optimization of SVR Kernel Function and Parameters223
7.3.3 Hybrid EO-Based Optimization for SVR Kernel Function and Parameters224
7.3.3.1 Chromosome Structure224
7.3.3.2 Fitness Function225
7.3.3.3 EO-SVR Workflow 226
7.3.4 Experimental Results 228
7.3.4.1 Approximation of Single-Variable Function228
7.3.4.2 Approximation of Multivariable Function233
7.4 Nonlinear Model Predictive Control with MA-EO238
7.4.1 Problem Formulation for NMPC Based on SVM Model239
7.4.2 Real-Time NMPC with SVM and EO-SQP242
7.4.2.1 Workflow of Proposed NMPC242
7.4.2.2 Encoding Strategy244
7.4.2.3 Selection of the Initial Population246
7.4.2.4 Termination Criteria of the NLP Solver246
7.4.2.5 Horizon-Based EO Mutation for NMPC Online Optimization 246
7.4.3 Simulation Studies247
7.5 Intelligent PID Control with Binary-Coded EO252
7.5.1 PID Controllers and Performance Indices252
7.5.2 BCEO Algorithm255
7.5.3 Experimental Results258
7.5.3.1 Single-Variable Controlled Plant258
7.5.3.2 Multivariable Controlled Plant261
7.5.3.3 Parameters and Control Performances267
7.6 Summary267
8 EO for Production Planning and Scheduling271
8.1 Introduction271
8.1.1 An Overview of HSM Scheduling272
8.1.2 Production Process273
8.1.3 Scheduling Objectives and Constraints274
8.2 Problem Formulation276
8.3 Hybrid Evolutionary Solutions with the Integration of GA and EO 280
8.3.1 Global Search Algorithm: Modified GA for Order Selection and Sequencing 280
8.3.1.1 Representation of Solutions 280
8.3.1.2 Population Initialization281
8.3.1.3 Fitness Function282
8.3.1.4 Genetic Operators282
8.3.2 Local Improving Algorithm: τ-EO285
8.3.2.1 Introduction to EO 286
8.3.2.2 EO for Improving the Body Section286
8.3.2.3 Hybrid Evolutionary Algorithms288
8.3.3 Design of a HSM-Scheduling System291
8.3.4 Computational Results293
8.4 Summary296
References297
Author Index315
Subject Index323
猜您喜欢
