Browse Books

Reviewer: Soubhik Chakraborty

The term "evolutionary algorithm" (EA) refers to a special category of "direct, probabilistic search and optimization algorithms (OA) gleaned from the model of organic evolution" [1]. They primarily include genetic algorithms (GAs), evolution strategies (ESs), and evolutionary programming (EP), all of which developed, interestingly, independent of one another. The EAs form an interdisciplinary paradigm with applications in biology, artificial intelligence (AI), and numerical optimization, not to mention engineering. This book will go a long way in meeting the requirements of a textbook cum reference material on the topic. The target readership includes, but is not limited to, students, teachers, researchers, and practitioners of computer science (CS) and operations research (OR), and, in fact, anyone who deals with evolutionary computation. Chapter 1, an introduction, familiarizes readers with the basic concepts of single and multiobjective (MO) optimization problems, after a helpful review of some mathematical preliminaries such as convex sets. The major topics covered include: multiobjective evolutionary algorithms (MOEAs), the Pareto epsilon model, a general OA overview, and EA basics; origins, mathematical foundations, and applications of MO optimization; and classifying techniques and the use of EA. Chapter 2 is on multiobjective problem (MOP) EA approaches. Here, the authors carefully explain the MOEA techniques, the structures of various MOEAs (such as nondominated sorting genetic algorithms (NSGA) and the Pareto envelope-based selection algorithm (PESA)), constraint handling strategies, critical MOEA elements (MOEA chromosomal representation), and MOEA design recapitulation. Chapter 3 covers MOEA local search and coevolution. This chapter is a new addition to the previous edition, and the authors cover intriguing topics like hybrid MOEA techniques, MOEA coevolutionary techniques, and coevolution and symbiosis in EAs and MOEAs, with quite lucid applications. Chapter 4 is on MOEA test suites; it broadly covers issues related to MOEA test function suites, MOP domain feature classification, generic scalable MOP test problems, and combinatorial and real-world MOEA test functions. Chapter 5 deals with MOEA testing and analysis, where experimental strategies, including a statistical approach, are discussed. Chapter 6 covers MOEA theory and issues like MOEA convergence, fitness functions, Pareto ranking, MOEA complexity and stability, and the interesting no free lunch (NFL) theorem. Chapter 7 should excite any applied scientist, as it is exclusively application oriented with MOEA applications to engineering (electrical and electronics, telecommunication, aeronautical, robotics, and control engineering), science (chemistry, physics, medicine, and CS), and industry (design, manufacture, scheduling, and management). Chapter 8, on MOEA parallelization, will interest the technically minded reader: the parallel MOEA (pMOEA) paradigm and models, such as the Island pMOEA model, are discussed. Chapter 9 is on multicriteria decision making, including fuzzy logic applications. Chapter 10, "Alternative Metaheuristics," targets areas like simulated annealing, tabu and scatter searches, the Ant system, distributed reinforcement learning, particle swarm optimization, and cultural algorithms. Having specialized in OR, I cherished reviewing such a fine book. The "Further Explorations" sections found at the end of each chapter?where suggested reading, exercises, and potential research areas are provided?are a strength of the book. Another strength is that it provides links to a complete set of teaching material, including tutorials. Also, the references are quite extensive and up to date. Despite all of the positive points, this book is not completely flawless. The omission of the vendor selection problem?an important multicriteria decision-making problem?in chapter 9 is a major lapse. There should have been some material on projection pursuit in chapter 1; an appendix for elementary probability, fuzzy logic, and genetics; and a glossary of mathematical symbols¿necessary for this sort of book. Also, the authors should have included more numerical examples to illustrate the theory; hopefully, they will do so in the next edition. Online Computing Reviews Service