Brief introduction to genetic algorithms
Marian B. Gorzałczany · Studies in fuzziness and soft computing · 2002
Among the four main classes of evolution-like and population-oriented methods of evolutionary computations [48, 60, 62, 138, 139, 172, 173, 236], that is, genetic algorithms, evolution strategies, evolutionary programming, and genetic programming, the first class plays a particularly important role. Genetic algorithms are a popular and widely used globalsearch paradigm based on principles imitating mechanisms of genetics, natural selection, evolution and heredity, including the evolutionary principle of survival of the fittest (to environment) individuals and extinction of the worst adapted individuals. The underlying principles of genetic algorithms were first formulated by Holland [138]. The mathematical framework was developed in the 1960s and was presented in his pioneering book [139]. An essential feature of the genetic-algorithmbased global searching of the solution domain is preserving the best possible balance between the two opposite requirements, that is, the use of the already-found best solutions and a possibly wide search of the solution domain. Genetic algorithms offer a compromise methodology, which eliminates many shortcomings of the two extreme approaches: traditional optimization techniques and random search methods. These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.