Computational Intelligence for Optimization

The field of optimization is interdisciplinary in nature, and has been making a significant impact on many disciplines. As a result, it is an indispensable tool for many practitioners in various fields. Conventional optimization techniques have been well established and widely published in many excellent textbooks. However, there are new techniques, such as neural networks, simulated anneal ing, stochastic machines, mean field theory, and genetic algorithms, which have been proven to be effective in solving global optimization problems. This book is intended to provide a technical description on the state-of-the-art development in advanced optimization techniques, specifically heuristic search, neural networks, simulated annealing, stochastic machines, mean field theory, and genetic algorithms, with emphasis on mathematical theory, implementa tion, and practical applications. The text is suitable for a first-year graduate course in electrical and computer engineering, computer science, and opera tional research programs. It may also be used as a reference for practicing engineers, scientists, operational researchers, and other specialists. This book is an outgrowth of a couple of special topic courses that we have been teaching for the past five years. In addition, it includes many results from our inter disciplinary research on the topic. The aforementioned advanced optimization techniques have received increasing attention over the last decade, but relatively few books have been produced.

Klappentext

The field of optimization is interdisciplinary in nature, and has made a significant impact on many areas of technology. As a result, optimization is an indispensable tool for many practitioners in various fields. Conventional optimization techniques are well established and widely published in many excellent textbooks. However, there are new techniques, such as simulated annealing, mean field theory, and genetic algorithms, which have proven to be effective in solving global optimization problems. Computational Intelligence for Optimization is intended as a technical description of the state-of-the-art developments in advanced optimization techniques, specifically simulated annealing, mean field theory, and genetic algorithms, with emphasis on mathematical theory, implementation, and practical applications. Most of the theories covered in this work and their applications are widely scattered in journals, technical reports, and conference proceedings of various fields, making it difficult for people new in the field to find easily. The authors have brought together a comprehensive and organized treatment of these techniques, thus filling a gap in the scientific literature. Computational Intelligence for Optimization is suitable for first-year graduate courses in electrical and computer engineering, computer science, and operational research programs. It may also be used as a reference work for practising engineers, scientists, operational researchers and other specialists.

Inhalt
1 Introduction.- 1.1 Computational Complexity.- 1.2 Survey of Optimization Techniques.- 1.3 Organization of the Book.- 1.4 Exploratory Problems.- 2 Heuristic Search Methods.- 2.1 Graph Search Algorithm.- 2.2 Heuristic Functions.- 2.3 A* Search Algorithm.- 2.4 Exploratory Problems.- 3 Hopfield Neural Networks.- 3.1 Discrete Hopfield Net.- 3.2 Continuous Hopfield Net.- 3.3 Content-Addressable Memory.- 3.4 Combinatorial Optimization.- 3.5 Exploratory Problems.- 4 Simulated Annealing and Stochastic Machines.- 4.1 Statistical Mechanics and The Metropolis Algorithm.- 4.2 Simulated Annealing.- 4.3 Stochastic Machines.- 4.4 Exploratory Problems.- 5 Mean Field Annealing.- 5.1 Mean Field Approximation.- 5.2 Saddle-Point Expansion.- 5.3 Stability.- 5.4 Parameters of the Mean Field Net.- 5.5 Graph Bipartition An Example.- 5.6 Exploratory Problems.- 6 Genetic Algorithms.- 6.1 Simple genetic Operators.- 6.2 An Illustrative Example.- 6.3 Why Do Genetic Algorithms Work?.- 6.4 Other Genetic Operators.- 6.5 Exploratory Problems.- 7 The Traveling Salesman Problem.- 7.1 Why Does the Hopfield Net Frequently Fail to Produce Valid Solutions?.- 7.2 Solving the TSP with Heuristic Search Algorithms.- 7.3 Solving the TSP with Simulated Annealing.- 7.4 Solving the TSP with Genetic Algorithms.- 7.5 An Overview of Eigenvalue Analysis.- 7.6 Derivation of ?1 of the Connection Matrix.- 7.7 Exploratory Problems.- 8 Telecommunications.- 8.1 Satellite Broadcast Scheduling.- 8.2 Maximizing Data Throughput in An Integrated TDMA Communications System.- 8.3 Summary.- 8.4 Exploratory Problems.- 9 Point Pattern Matching.- 9.1 Problem Formulation.- 9.2 The Simulated Annealing Framework.- 9.3 Evolutionary Programming.- 9.4 Summary.- 9.5 Exploratory Problems.- 10 Multiprocessor Scheduling.- 10.1 Model andDefinitions.- 10.2 Mean Field Annealing.- 10.3 Genetic Algorithm.- 10.4 Exploratory Problems.- 11 Job Shop Scheduling.- 11.1 Types of Schedules.- 11.2 A Genetic Algorithm for JSP.- 11.3 Simulation Results.- 11.4 Exploratory Problems.- References.