Representation for Genetic and Evolutionary Algorithms: Studies in Fuzziness and Soft Computing

Citation

Rigelsford, J. (2003), "Representation for Genetic and Evolutionary Algorithms: Studies in Fuzziness and Soft Computing", Industrial Robot, Vol. 30 No. 2. https://doi.org/10.1108/ir.2003.04930bae.002

Representation for Genetic and Evolutionary Algorithms: Studies in Fuzziness and Soft Computing

Representation for Genetic and Evolutionary Algorithms: Studies in Fuzziness and Soft Computing

F. RothlaufPhysica-Verlag2002289 pp.ISBN 3-7908-1496-2£45.50 hardback

Keywords: Algorithms

This book investigates how the performance of genetic and evolutionary algorithms (GEAs) is influenced by their representations. It aims to provide an applicable representation theory that can assist researchers and practitioners in designing the proper representation for their problem.

After a brief introduction, chapter 2 discusses Representation for Genetic and Evolutionary Algorithms. Topics addressed include genetic representations, genetic and evolutionary algorithms, problem difficulty, and existing recommendations for the design of efficient representations of GEAs.

Chapters 3 and 4 present Three Elements of a Theory of Genetic and Evolutionary Representations, and Time-Quality Framework for a Theory-Based Analysis and Design of Representations, respectively. Amongst the subjects discussed are redundancy, building block-scaling, distance distortion, elements of the framework, implications for the design representations, and solution quality and time to convergence.

Two integer optimisation problems, binary string representations, a theoretical comparison and empirical results, are amongst the topics discussed in chapter 5, Analysis of Binary Representations of Integers. Chapter 6, Analysis of Tree Representations, discusses the tree design problem, Prüfer numbers, link and node biased encoding, and characteristic vector encoding, while chapter 7 addresses the Design of Tree Representations. This section focuses on network random keys (NetKeys) and a direct tree representation (NetDir).

Chapter 8 addresses the Performance of Genetic and Evolutionary Algorithms on Tree Problems, and discusses GEA performance on scalable test tree problems, and GEA performance on the optimal communication spanning tree problem. The final chapter of the book provides a summary, conclusions and areas of interest for future work. An appendix is included providing Palmer’s Raidl’s and Berry’s test instances for optimal communications spanning tree problems.

Overall, Representations for Genetic and Evolutionary Algorithms is both informative and enjoyable to read. A sound mathematical background is required to understand the theoretical models described in the book but the author does a good job to ensure that they are not daunting. The book is suitable for those with experience in GEA design and for the newcomer.