You are previewing Multi-Objective Optimization in Computational Intelligence: Theory and Practice.
O'Reilly logo
Multi-Objective Optimization in Computational Intelligence: Theory and Practice

Book Description

"Multi-objective optimization (MO) is a fast-developing field in computational intelligence research. Giving decision makers more options to choose from using some post-analysis preference information, there are a number of competitive MO techniques with an increasingly large number of MO real-world applications.

Multi-Objective Optimization in Computational Intelligence: Theory and Practice explores the theoretical, as well as empirical, performance of MOs on a wide range of optimization issues including combinatorial, real-valued, dynamic, and noisy problems. This book provides scholars, academics, and practitioners with a fundamental, comprehensive collection of research on multi-objective optimization techniques, applications, and practices."

Table of Contents

  1. Copyright
  2. Reviewer List
  3. Foreword
  4. Preface
  5. Acknowledgment
  6. I. Fundamentals
    1. I. An Introduction to Multi-Objective Optimization
      1. ABSTRACT
      2. OVERVIEW
      3. CONCEPTS AND NOTATIONS
      4. TRADITIONAL MULTI-OBJECTIVE ALGORITHMS
        1. No-Preference Methods
        2. Posteriori Methods
        3. Priori Methods
        4. Interactive Methods
      5. MULTI-OBJECTIVE EVOLUTIONARY ALGORITHMS
        1. Overview
        2. Non-Elitism Approach
        3. Elitism Approach
        4. Selected MOEAs
          1. Nondominated Sorting Genetic Algorithms Version 2: NSGA-II
          2. A Pareto-Frontier Differential Evolution Algorithm for MOPs: PDE
          3. Strength Pareto Evolutionary Algorithm 2: SPEA2
          4. Pareto Archived Evolutionary Strategy: PAES
          5. Multi-Objective Particle Swarm Optimizer: MOPSO
      6. PROBLEM DIFFICULTIES AND RESEARCH ISSUES
      7. PERFORMANCE ASSESSMENTS
        1. Metrics Evaluating Closeness to the POF
        2. Metrics Evaluating Diversity Among Obtained Nondominated Solutions
        3. Metrics Evaluating Both Closeness and Diversity
        4. Statistical Testing
      8. CONCLUSION
      9. REFERENCES
    2. II. Multi-Objective Particles Swarm Optimization Approaches
      1. ABSTRACT
      2. INTRODUCTION
      3. BACKGROUND MATERIAL
        1. Basic Multi-Objective Optimization Concepts
        2. Particle Swarm Optimization
      4. KEY CONCEPTS OF MULTI-OBJECTIVE PSO ALGORITHMS
      5. ESTABLISHED MULTI-OBJECTIVE PSO APPROACHES
        1. Algorithms that Exploit Each Objective Function Separately
        2. Objective Function Aggregation Approaches
        3. Objective Function Ordering Approaches
        4. Non-Pareto, Vector Evaluated Approaches
        5. Algorithms Based on Pareto Dominance
      6. FUTURE RESEARCH DIRECTIONS
      7. CONCLUSION
      8. REFERENCES
      9. ADDITIONAL READING
        1. Book Chapters
        2. Journal Papers
        3. Conference Papers
        4. Theses
    3. III. Generalized Differential Evolution for Constrained Multi-Objective Optimization
      1. ABSTRACT
      2. INTRODUCTION
      3. BACKGROUND
        1. Multi-Objective Optimization with Constraints
        2. Differential Evolution
        3. Basic Differential Evolution, DE/rand/1/bin
            1. Initialization of Population
            2. Mutation and Crossover
            3. Selection
            4. Overall Algorithm
          1. Differential Evolution for Multiple Objectives and Constraints
      4. GENERALIZED DIFFERENTIAL EVOLUTION
        1. The First Version, GDE1
        2. The Second Version, GDE2
        3. The Third Version, GDE3
          1. Diversity Maintenance for Bi-Objective Problems
          2. Diversity Maintenance for Many-Objective Problems
        4. Study of Control Parameter Values for GDE
        5. Constrained Optimization with the GDE Versions
      5. FUTURE RESEARCH DIRECTIONS
      6. CONCLUSION
      7. REFERENCES
      8. ADDITIONAL READING
    4. IV. Towards a More Efficient Multi-Objective Particle Swarm Optimizer
      1. ABSTRACT
      2. INTRODUCTION
      3. BASIC CONCEPTS
        1. Multi-Objective Optimization
          1. Pareto Dominance
          2. Pareto Front
        2. Particle Swarm Optimization (PSO)
          1. Multi-Objective Algorithms Based on PSO
        3. Scatter Search
        4. Handling Well-Distributed Solutions
      4. OUR PROPOSED APPROACH
        1. Phase I: Particle Swarm Optimization
        2. Phase II: Scatter Search
        3. Use of ε-Dominance
      5. COMPARISON OF RESULTS
        1. Discussion of Results
          1. C-MOPSOSS vs NSGA-II Comparison
          2. Comparison Between Our C-MOPSOSS and MOPSO
          3. Comparing the Two Stages of Our Approach
          4. Convergence Capability of Our C-MOPSOSS
      6. CONCLUSION
      7. FUTURE RESEARCH DIRECTIONS
      8. REFERENCES
      9. ADDITIONAL READING
      10. ENDNOTES
    5. V. Multi-Objective Optimization Using Artificial Immune Systems
      1. ABSTRACT
      2. INTRODUCTION
      3. RELATED BACKGROUND
        1. Antigen Ψ
        2. Antibody and Antibody Population
        3. Ab-Ag Affinity ∂
        4. Ab-Ab Affinity Ω
        5. Dominant Antibody
      4. IMMUNE DOMINANCE CLONAL MULTI-OBJECTIVE ALGORITHM
        1. Description of the Algorithm
          1. Immune Dominance Recognizing Operation
            1. Algorithm 1: Immune Dominance Recognizing Operation
          2. Immune Dominance Clone Operation
            1. Algorithm 2: Immune Dominance Clone Operation
          3. Dominance Clonal Selection Operation
            1. Algorithm 3: Dominance Clonal Selection Operation
          4. The Main Loop
            1. Algorithm 4: Immune Dominance Clonal Multi-objective Algorithm, IDCMA
        2. Analysis of the Algorithm
          1. Fitness Assignment and Population Evolution
          2. Population Diversity
          3. Computational Complexity
        3. Experimental Study on Combinatorial MO Problems
      5. NONDOMINATED NEIGHBOR IMMUNE ALGORITHM
        1. Description of the Algorithm
          1. Proportional Cloning
          2. Recombination and Hypermutation
          3. Fitness Assignment and Population Evolution
          4. Computational Complexity
        2. Evaluation of NNIA'S Effectiveness
          1. Experimental Setup
          2. Comparison of NNIA with PESA-II, SPEA2, NSGA-II and MISA
          3. Comparison of NNIA with and without Recombination
      6. CONCLUDING REMARKS
      7. FUTURE RESEARCH DIRECTIONS
      8. ACKNOWLEDGMENT
      9. REFERENCES
      10. ADDITIONAL READING
    6. VI. Lexicographic Goal Programming and Assessment Tools for a Combinatorial Production Problem
      1. ABSTRACT
      2. INTRODUCTION
      3. LITERATURE REVIEW
      4. MODELING THE MULTI-CRITERIA PRODUCTION PROBLEM
      5. LEXICOGRAPHIC GOAL PROGRAMMING FOR USE IN SOLUTION METHODOLOGY EVALUATION
      6. ASSESSMENT TOOLS
        1. Graphical Analysis Tools
        2. Efficacy Index Equations
        3. Statistical Regression
        4. Statistical Coefficient of Determination
        5. Performance Assessment Experimental Benchmark Data Set
      7. H-K HEURISTIC
        1. Heuristic Search Background
        2. Heuristic Motivation and Introduction
        3. The H-K Process and DLBP Application
      8. GENETIC ALGORITHM
        1. Ga Model Description
        2. DLBP-Specific Genetic Algorithm Architecture
        3. DLBP-Specific GA Qualitative Modifications
        4. DLBP-Specific GA Quantitative Modifications
      9. NUMERICAL COMPARISON AND QUALITATIVE ANALYSIS
        1. Table-Based Qualitative Assessment
        2. Graph-Based Qualitative Assessment and Quantitative Analysis Using Efficacy Indices and Regression
      10. FUTURE DIRECTIONS
      11. CONCLUSION
      12. REFERENCES
      13. ADDITIONAL READING
    7. VII. Evolutionary Population Dynamics and Multi-Objective Optimisation Problems
      1. ABSTRACT
      2. INTRODUCTION
      3. MULTI-OBJECTIVE ORIENTED METAHEURISTICS
        1. Particle Swarm Optimisation
        2. Ant Colony Optimisation
        3. Extremal Optimisation
      4. EVOLUTIONARY POPULATION DYNAMICS
      5. EPSOC
        1. Applying EPD
      6. MOPSO AND EPD
      7. COMPUTATIONAL EXAMPLE
      8. CONCLUDING REMARKS AND FUTURE RESEARCH DIRECTIONS
      9. ACKNOWLEDGMENT
      10. REFERENCES
      11. ADDITIONAL READING
      12. ENDNOTES
  7. II. Applications
    1. VIII. Multi-Objective Evolutionary Algorithms for Sensor Network Design
      1. ABSTRACT
      2. INTRODUCTION
      3. BACKGROUND ON SENSOR NETWORK DESIGN
      4. MULTI-OBJECTIVE OPTIMIZATION PROBLEMS IN SENSOR NETWORKS
        1. Sensor Placement Problem
          1. Detection Probability
          2. Deployment Cost
          3. Energy Consumption
          4. Related Work
        2. Mobile Agent Routing
          1. Detection Accuracy
          2. Path Loss
          3. Energy Consumption
          4. Related Work
        3. Data Aggregation
          1. Data Accuracy
          2. Network Lifetime
          3. Latency
          4. Related Work
        4. Area Coverage
          1. Related Work
      5. PRIOR WORK: MULTIPLE OBJECTIVES IN SENSOR NETWORK DESIGN
      6. MULTI-OBJECTIVE OPTIMIZATION ALGORITHMS
        1. Weight Based Genetic Algorithm
        2. Multi-Objective Evolutionary Algorithms (MOEAs)
          1. Description of NSGA-II
          2. Description of EMOCA
      7. MOBILE AGENT ROUTING IN SENSOR NETWORKS
        1. Mobile Agent Distributed Sensor Network (MADSN) Architecture
        2. Problem Formulation
          1. Energy Consumption
          2. Path Loss
          3. Detection Accuracy
        3. Problem Representation and Genetic Operators
          1. Representation
          2. Crossover
      8. SENSOR PLACEMENT FOR ENERGY EFFICIENT TARGET DETECTION
        1. Problem Formulation
        2. Probability of Target Detection
          1. Data Level Fusion Model: Calculation of Л(S,T,i,j)
          2. Decision Fusion Model: Calculation of Л(S,T,i,j)
        3. Energy Cost Analysis
          1. Data Level Fusion Model: Calculation of Total Energy
          2. Decision Fusion Model: Calculation of Total Energy
        4. Problem Representation and Genetic Operators
        5. Simulation Results
      9. POSTPROCESSING OF SOLUTIONS OBTAINED BY A MOO ALGORITHM
      10. FUTURE RESEARCH DIRECTIONS
      11. CONCLUSION
      12. REFERENCES
      13. ADDITIONAL READING
        1. Multi-Objective Evolutionary Algorithms
        2. Sensor Networks
    2. IX. Evolutionary Multi-Objective Optimization for DNA Sequence Design
      1. ABSTRACT
      2. INTRODUCTION
      3. BACKGROUND
        1. Criteria of Good DNA Sequence
        2. DNA Computing Sequence Design
        3. Microarray Probe Design
        4. Multiplex PCR Primer Design
      4. ε-MULTI-OBJECTIVE EVOLUTIONARY ALGORITHM
      5. DNA SEQUENCE DESIGN FOR DNA COMPUTING
      6. MULTI-OBJECTIVE FORMULATION OF DNA COMPUTING SEQUENCE DESIGN
        1. Multi-Objective Evolutionary Sequence Optimization
        2. DNA Computing Sequence Optimization Results
      7. PROBE DESIGN FOR OLIGONUCLEOTIDE MICROARRAY
        1. Multi-Objective Formulation of Oligonucleotide Microarray Probe Design
        2. Multi-Objective Evolutionary Probe Optimization
        3. Probe Selection Results
      8. PRIMER DESIGN FOR MULTIPLEX POLYMERASE CHAIN REACTION
        1. Multi-Objective Formulation of Multiplex PCR Assay
        2. Hybrid Multi-Objective Evolutionary Design for Multiplex PCR Assay
        3. Multiplex PCR Assay Selection Results
      9. SUMMARY
      10. FUTURE RESEARCH DIRECTIONS
      11. ACKNOWLEDGMENT
      12. REFERENCES
      13. ADDITIONAL READING
        1. For General Bioinformatics Backgrounds
        2. For DNA Computing
        3. For Probe Selection
        4. For Primer Selection
    3. X. Computational Intelligence to Speed-Up Multi-Objective Design Space Exploration of Embedded Systems
      1. ABSTRACT
      2. INTRODUCTION
      3. BACKGROUND
        1. Design Space Exploration Approaches
        2. Fast Evaluation through Approximated Models
        3. Coupling Efficient Exploration with Fast System Evaluation
        4. Contribution
      4. FORMULATION OF THE PROBLEM
      5. THE MOGA+FUZZY APPROACH TO SPEED-UP DESIGN SPACE EXPLORATION
      6. MULTI-OBJECTIVE EVOLUTIONARY ALGORITHM
      7. FUZZY FUNCTION APPROXIMATION
      8. SIMULATION FRAMEWORK AND QUALITY MEASURES
        1. Parameterized System Architecture
        2. Simulation Flow
        3. The Hierarchical Fuzzy Model
        4. Assessment of Pareto Set Approximations
      9. EXPERIMENTS AND RESULTS
        1. A Case Study: Application Specific Cache Customization
        2. DSE of a Parameterized VLIW-Based Embedded System
      10. FUTURE RESEARCH DIRECTIONS
      11. CONCLUSION
      12. REFERENCES
      13. ADDITIONAL READINGS
      14. ENDNOTE
    4. A. APPENDIX: ALGORITHMS PSEUDO-CODE
        1. Algorithm 1. SPEA2 Pseudo-Code
        2. Algorithm 2. ParEGO Pseudo-Code
    5. XI. Walking with EMO: Multi-Objective Robotics for Evolving Two, Four, and Six-Legged Locomotion
      1. ABSTRACT
      2. INTRODUCTION
      3. PREVIOUS WORKS ON EVOLUTIONARY ROBOTICS AND MULTI-OBJECTIVE OPTIMIZATION
      4. ATIFICIAL EVOLUTION AND VIRTUAL SIMULATION SETUP
        1. Physics Engine
        2. Simulated Robot Morphologies
        3. Evolutionary Objectives
        4. The Evolutionary Neural Network Algorithm and Controller
        5. Genotype Representation
      5. EMO-BASED GENERATION OF 2, 4 AND 6-LEGGED LOCOMOTION
        1. SPANN-R Algorithm for Multi-Objective Evolution of Robotic Controllers
          1. Experimental Setup
          2. Evolved Pareto Controllers for Two-Legged Locomotion
          3. Evolved Pareto Controllers for Four-Legged Locomotion
          4. Evolved Pareto Controllers for Six-Legged Locomotion
        2. Complexity Comparison Between Bipedal, Quadrupedal and Hexapedal Controllers
      6. OPERATIONAL AND LIMB DYNAMICS
        1. Operational Dynamics under Noisy Conditions
        2. Operational Dynamics Beyond the Evolutionary Window
        3. Limb Dynamics
      7. EMPIRICAL COMPARISON OF DIFFERENT EMO ALGORITHMS
        1. SPANN-R vs. NSGA-II for Pareto Evolution of Locomotion Controllers
      8. CONCLUSION
      9. FUTURE RESEARCH DIRECTIONS
      10. REFERENCES
      11. ADDITIONAL READINGS
        1. Multi-Objective Robotics
        2. Evolutionary Algorithms
        3. Evolutionary Multi-Objective Optimization
        4. Evolutionary Robotics
        5. Physics-Based Simulators
        6. Conferences of Interest
    6. XII. Evolutionary Multi-Objective Optimization in Energy Conversion Systems: From Component Detail to System Configuration
      1. ABSTRACT
      2. INTRODUCTION
      3. DESIGN OF COMPONENT DETAILS
        1. Optimization of Aerofoil Shape for Axial Compressor Blades
        2. Optimal Design of Heat Exchangers
        3. Design of Optimal Rotors for Horizontal-Axis Wind Turbines
      4. DESIGN OF ENERGY SYSTEMS
        1. Optimization of System Design Parameters
        2. Optimal Synthesis of Heat Exchanger Networks
        3. Optimal Synthesis and Parameter Design of Complex Energy Systems
        4. Design of Optimized District Heating Networks
      5. OPERATION OF ENERGY SYSTEMS
        1. Optimized Operation of a Control System
        2. Diagnosis of Malfunction with Fuzzy Expert Systems
      6. CONCLUSION
      7. FUTURE RESEARCH DIRECTIONS
      8. ACKNOWLEDGMENT
      9. REFERENCES
      10. ADDITIONAL READINGS
        1. About Component Design
        2. About Energy System Design
        3. About Energy System Operation
  8. XIII. Evolutionary Multi-Objective Optimization for Assignment Problems*
    1. ABSTRACT
    2. INTRODUCTION
    3. MULTI-OBJECTIVE ASSIGNMENT PROBLEMS
      1. Linear Gate Assignment Problem
      2. Multi-Objective Quadratic Assignment problem
      3. Airman Assignment Problem
      4. Causality Assignment Problem
      5. Fixed Channel Assignment Problem
      6. Frequency Assignment Problems
      7. Multilevel Generalized Assignment Problem
      8. Resource Allocation Problem
      9. Other Multi-Objective Assignment Problem Instances
    4. MOEA OPERATORS FOR ASSIGNMENT PROBLEMS
      1. Chromosome Representations
        1. Variable Length Chromosomes
        2. Fixed Length Chromosomes
      2. Mutation Operators
      3. Crossover Operators
      4. Selection Methods
        1. Selection for Recombination
        2. Generational Selection
    5. SUMMARY
    6. REFERENCES
    7. FUTURE RESEARCH DIRECTIONS
    8. ADDITIONAL READING
    9. ENDNOTE
  9. XIV. Evolutionary Multi-Objective Optimization in Military Applications*
    1. ABSTRACT
    2. INTRODUCTION
    3. COMMUNICATION NETWORKS
      1. Network Design
      2. Network Routing
        1. Multi-Objective Mobile Agent Routing in Wireless Sensor Networks
      3. Network Sensor Layout
        1. Layout Optimization for a Wireless Sensor Network
        2. Automated Placement of Wireless Sensor Network Nodes
        3. Resource Management in Wideband CDMA Systems
    4. MANAGEMENT OF MILITARY RESOURCES
      1. Lifetime Management of Military Platforms
      2. Military Aircraft Engine Maintenance Scheduling
    5. MISSION PLANNING
      1. Planning
      2. Courses of Action Planning
      3. Mission Planning and Routing
    6. PERSONNEL ASSIGNMENT
      1. Sailor Assignment Problem
      2. Airman Assignment Problem
    7. MOEA INTEGRATED MILITARY SIMULATION
    8. DESIGN OF INNOVATIVE EQUIPMENT WITH MOEAS
      1. Low-Power Laser Design
      2. Autopilot Controller
    9. OTHER APPLICATIONS
      1. Groundwater Remediation (Knarr, 2003) (Singh & Minsker, 2004)
      2. Optimization of UAV Communications (Kleeman, 2004) (Day, 2005)
      3. Many-Objective Radar Waveform Optimization (Hughes, 2007)
    10. SUMMARY
    11. REFERENCES
    12. FUTURE RESEARCH DIRECTIONS
    13. ADDITIONAL READING (MOEA MILITARY RELEVANCE)
    14. ENDNOTE
  10. Compilation of References
  11. About the Contributors