Advances in evolutionary computing. Theory and applications (Q1871716)

Without coming into explanatory details that would remain inevitably inconsistent and incomplete, we expose directly the following abbreviations proposed by the editors of the book for the notions to be subsequently: GA = Genetic Algorithm; GP = Genetic Programming; EA = Evolutionary Algorithm; ES = Evolutionary Strategies; EP = Evolutionary Programming. The term Evolutionary Computing (EC hereafter) would comprise the foundations and applications of certain heuristic techniques based on the principles of natural evolution, and whose material object study should be the Evolutionary Algorithms (EAs), understood as computing devices that try to mimic the processes taking place in natural evolution. One can appreciate that \(\text{EC}= \text{GAs}\cup \text{GP}\cup \text{ES}\cup \text{EP}\). The present book offers a collection of 40 high-quality articles (chapters) on EC, grouped in two parts: (I) the theoretical aspects of EC (23 articles), and (II) applicational issues demonstrating the usefulness/success of EC methodologies (17 articles). Here there are some brief notes of the editors on each chapter (article) contents. In Part I (Theory), Chapter 1 (Smoothness, Ruggedness and Neutrality of Fitness Landscapes) studies the structures of fitness landscapes to provide a suitable mathematical framework for investigating the evolvability of complex systems. Chapter 2 (Fast EAs) tries to demonstrate the role of the search step in approximating the landscape by using different hybrid EAs. Chapter 3 (Visualizing EC) surveys the existing work on visualizing EAs, identifying some key issues for future research. New schemes or operator design are described in Chapters 4-5. Chapter 4 (New Schemes of Biologically Inspired EC) proposes a parameter free GA (Genetic Algorithm), based on the disparity theory of evolution which exploits different mutation rates of evolution population sizes. Chapter 5 (On the Design of Problem specific EAs) suggests guidelines for the design of genetic operators and the representation of phenotypic space to solve specific types of problems. The application of this concept is illustrated by a systematic design of a GP system for finding Boolean functions. This is the first GP system that has reportedly found die 12-parity function. Chapter 6 (Multi-parent Recombination in EC) demonstrates the utility of multiparent reproduction with successful results. The traditional debate of mutation and crossover is also considered in the light of multiparent reproduction. Chapter 7 (TCG-2: A Test-case Generator for Nonlinear Parameter Optimization Techniques) proposes a test case generator that is capable to create various test problems with different characteristics, including the dimensionality of the problem, number of local optima, number of active constraints at the optimum, topology of the feasible search space etc. Such a test case generator is useful for analyzing and comparing different constraint handling techniques. Chapter 8 (A Real-coded GA Using the Unimodal Normal Distribution Crossover) investigates a new crossover operator which works efficiently for optimization problems with epistasis among parameters. Chapter 9 (Designing EAs for Dynamic Optimization Problems) provides a good survey of the EC literature for dynamic optimization problems. The article offers a classification of the same class of problems and suggests a new technique, for this task using a multi-population structure. Chapter 10 (Multi-objective EAs: Introducing Bias Among Pareto-optimal Solutions) considers some classical techniques to identify a preferred or compromise solution in complex multi-objective optimization problems. A novel biased sharing method is introduced to finding a biased distribution of Pareto-optimal solutions in multi-objective GAs. The utility of gene expression in scalable genetic search is studied in Chanter 11 (Gene Expression and Scalable Genetic Search). Chapter 12 (Solving Permutation Problems with the Ordering Messy GA) presents an ordering messy GA that is able to solve difficult permutation problems according to the experimental results. Chapter 13 (Effects of Adding Permutations to Phenotypic Parameters in GAs for searching Robust Solutions) investigates techniques of detecting GA-based robust solutions. The authors introduce a natural variation of their robust solution searching scheme (the new method being called sixth-evaluation model), where perturbations are given more than once for each evaluation of an individual, and provide a comparative study for die convergence properties of the methods. In Chapter 14 (Evolution of Strategies for Resource Protection Problems), EC is used to evolve finite-state machines that have an optimal number of states for improving the performance in resource allocation. Chapters 15 and 16 try to link EC with statistical inferencing. In Chapter 15 (A Unified Bayesian Framework for Evolutionary Learning and Optimization), EC is viewed as a Bayesian inference that iteratively updates the posterior distribution of a population from a prior knowledge and observation of new individuals, in order to find an individual with the maximum posterior probability. Chapter 16 (Designed Sampling with Crossover Operator) is an attempt to combine experimental design and EC into a single search strategy using a specific type of recombination function called a deterministic crossover operator. Theoretical understanding of biology and its simulation is the main topic of chapters 17-19. Chapter 17 (EC for Evolutionary Theory) aims to use EAs to reunite the theoretical and experimental biology. Chapter 18 (Computational Embryology) presents a brief survey on using embryology and genetics in developmental biology. The application of two embryological is illustrated. Chapter 19 (An Evolutionary Approach to Synthetic Biology) describes how to inoculate the process of natural evolution is an artificial medium, finds the natural form of the living organisms in the artificial medium, and suggests a possible means of harnessing the evolutionary process for the production of complex computer software. Chapters 20-22 deal with other heuristic algorithms closely related to EAs. The purpose of Chapter 20 (Scatter Search) is to demonstrate the development of scatter search procedures by illustrating their application to a class of nonlinear optimization problems of bounded variables. Key ideas for future research issues are highlighted. Chapter 21 (The Ant Colony Optimization Paradigm for Combinatorial Optimization) applies several ant colony optimization techniques on a number of hard optimization problems, with specific attention to a new algorithm called ANTS. Chapter 22 (Evolving Coordinated Agents) examines a co-evolutionary optimization for evolving agent groups in the framework of intelligent multiagent systems, and suggests an adaptive system approach. Chapter 23 (Exploring the Predictable) studies an embedded active learner that can be limit its prediction to arbitrary computable aspects of spatio-temporal events using probabilistic algorithms. Part II (Applications) of the book begins with Chapter 24 (Approaches to Combining Local and Evolutionary Search for Training Neural Networks), which demonstrates a method to combine local search with evolutionry search techniques for neural network learning, order to reduce the computational time. Chapters 25-27 apply EC techniques to analog circuit design. Chapter 25 (Evolving Analog Circuits by Variable Length Chromosomes) shows how to use variable-length GAs to evolve analog circuits. Chapter 26 (Human-competitive Applications of GP) recommends a method for applying GP techniques for due automatic synthesis of topology and sizing for analog electrical circuits, the synthesis of placement and routing for circuits, and the synthesis of both the topology and tuning of controllers. Chapter 27 (EAs for the Physical Design of VLSI Circuits) states an EC framework for the design of a physical problem, and discusses the important requirements of an enlarged acceptance of EC-based approaches within the VLSI community. The issue of designing communication channels is considered in Chapters 28-29. More precisely, Chapter 28 (An EA for the Antenna Placement Problem) reports the application of EAs constrained to multi-objective, large, real-world antena placement problems for mobile radio networks. Chapter 29 (Routing Optimization in Corporate Networks by EAs) shows how to find a routing table that increases the performance of a communication, non-hierarchical network. Chapters 30-32 discuss scheduling by EAs. Chapter 30 (GAs and Timetabling) presents a survey with a critical analysis on the application of EC in timetable scheduling problems, and the special need of a wide-ranging investigation of these problems. Chapter 31 (Machine Learning by Schedule Decomposition) describes techniques that use GAs as metaheuristics for guiding an optimal design schedule decomposition sequence, and solving with improved results the minimum makespan problem for job shop scheduling. Chapter 32 (Scheduling of Bus Drivers' Service by a GA) submits a scheme for driver scheduling along various routes, using heuristics and GAs. Data-mining is considered in Chapters 33-34. An excellent inspection of different EAs for data-mining problems is comprised in Chapter 33 (A Survey of EAs for Data Mining and Knowledge Discovery). Chapter 34 (Data Mining from Clinical Data Using Interactive EC) illustrates the use of interactivc EC and interactive features from (subsets of) clinical data for data-mining problems. Chapter 35 (Learning-integrated Interactive Image Segmentation) discusses a GA-guided approach to image segmentation, and suggests techniques for physics-based segmentation evaluation, novel crossover operator and fitness functions. Chapter 36 (An Immunogenetic Approach in Chemical Spectrum Recognition) proposes an immunogenetics approach to recognize composite spectra for chemical analyses, while Chapter 37 (Application of EC to Protein Folding) describes GA-based techniques for two-dimensional protein folding. Chapter 38 (Evolutionary Generation of Regrasping Motion) advises an EC method to control the regrasping notion of a four-fingered robot hand. Chapter 39 (Recent Trends in Learning Classifier Systems Research) reviews advances and trends in learning classifier systems. Finally, Chapter 40 (Evolving a Nearly Expert Checkers Player) illustrates a hybrid technique for exploiting the neural networks and EC for designing a computer program that plays checkers at an expert level.