Evolutionary algorithm

From Glossary

Jump to: navigation, search

This is a term that applies to any metaheuristic based on some biological metaphor that is a dynamical system whose evolution rules have a stochastic component. In particular, it includes genetic algorithms, scatter search, Particle Swarm Optimization, and some neural networks. It emerged as a unification of related computational methods, including cellular automata, which was conceived by Ulam and von Neumann in the 1940s. This unification is not universal. Some researchers cling to the GA metaphor but say that evolutionary algorithms have more operators; some say that the GA population is composed of bit strings, whereas evolutionary algorithms can have a broader range of population structures. It is the concept of evolution dynamics, which is population-based and stochastic, that distinguishes this class of algorithms from other methods of global optimization. Evolutionary computation uses the evolutionary algorithm as its foundation; typically, it is massively parallel.

Personal tools