×

zbMATH — the first resource for mathematics

Evolutionary computation: the fossil record. (English) Zbl 0908.68210
Piscataway, NJ: IEEE. xiii, 641 p. (1998).
Evolutionary computation is the field that studies the properties of the evolutionary algorithms. Over successive iteration an evolutionary algorithm can drive a population toward particular optima on a response surface that represents the measurable worth of each possible individual that might reside in a population. Natural evolution does not occur in discontinuous time intervals, but the use of a digital computer requires discrete events. Therefore, the process of evolution can be modeled algorithmically and simulated on a computer e.g. by means of difference equations. The most elementary model may be summarized as the difference equation \(x[t+1]= s(v(x[t]))\), where the population at time \(t\), denoted as \(x[t]\), is operated on by random variation \(v\), and selection \(s\), to give rise to a new population \(x[t+1]\) at time \(t+1\).
Although the term evolutionary computation was invented as recently as 1991, the field has a history that spans four decades. Many independent efforts to simulate evolution on a computer were offered in the 1950s and 1960s.
This volume of “the fossil record” includes 23 chapters. Each chapter consist of an introductory text and one or more reprinted papers of different authors. With the exception of the information in the first chapter, the fossil record presented here begins in 1950s and progresses forward through time, concluding with more contemporary research. This chronology reveals the multiple independent efforts in evolutionary computation in light of the state of-the-art at time those contributions were offered.
The papers selected for reprinting met one or more of the following criteria: (1) They offered a first or very early attempt at specific approach, (2) they had a significant impact on the future development of the field, (3) they could have had a significant impact had they received due attention, (4) they represented a key turning point in the field, or (5) the editor found them to be personally interesting. The topics founded important by the editor of the volume are best illustrated by the chapter’s titles which are as follow (in parentheses: number of reprinted papers).
Chapter 1: An introduction to evolutionary computation (2), Chapter 2: Evolving control circuits for autonomous robots (1), Chapter 3: Simulating the evolution of genetic systems (3), Chapter 4: Evolving online productivity (1) , Chapter 5: Evolving computer programs (2), Chapter 6: Artificial life and evolving strategies (2), Chapter 7: Artificial intelligence through simulated evolution (2), Chapter 8: Evolutionary experimentation (1), Chapter 9: Evolution and optimization (1), Chapter 10: Evolutionary algorithms for system identification (1), Chapter 11: Co-evolution, self-adaptation, and crossover (1), Chapter 12: Evolving populations (1), Chapter 13: Artificial ecosystems (1), Chapter 14: Soft selection (2), Chapter 15: Schema processing and the K-armed bandit (1), Chapter 16: Classifier systems (1), Chapter 17: Evolving neural networks (1), Chapter 18: Evolutionary computation and the travelling salesman problem (2), Chapter 19: The iterated prisoner’s dilemma (1), Chapter 20: Implicit parallelism and representations (1), Chapter 21: Fuzzy evolution (1), Chapter 22: Evolving programs using symbolic expressions (1), Chapter 23: Tierra and emergent properties (1), Items Epilogue, Author Index, Subject Index and about the Editor close the book.

MSC:
68U20 Simulation (MSC2010)
68-06 Proceedings, conferences, collections, etc. pertaining to computer science
68-03 History of computer science
01A75 Collected or selected works; reprintings or translations of classics
00B15 Collections of articles of miscellaneous specific interest
PDF BibTeX Cite