Evolutionary programming

Evolutionary programming is an evolutionary algorithm, where a share of new population is created by mutation of previous population without crossover.^[1][2] Evolutionary programming differs from evolution strategy ES(${\displaystyle \mu +\lambda }$) in one detail.^[1] All individuals are selected for the new population, while in ES(${\displaystyle \mu +\lambda }$), every individual has the same probability to be selected. It is one of the four major evolutionary algorithm paradigms.^[3]

History

It was first used by Lawrence J. Fogel in the US in 1960 in order to use simulated evolution as a learning process aiming to generate artificial intelligence.^[4] It was used to evolve finite-state machines as predictors.^[5]

Timeline of EP - selected algorithms^[1]

Year	Description	Reference
1966	EP introduced by Fogel et al.	[6]
1992	Improved fast EP - Cauchy mutation is used instead of Gaussian mutation	[7]
2002	Generalized EP - usage of Lévy-type mutation	[8]
2012	Diversity-guided EP - Mutation step size is guided by diversity	[9]
2013	Adaptive EP - The number of successful mutations determines the strategy parameter	[10]
2014	Social EP - Social cognitive model is applied meaning replacing individuals with cognitive agents	[11]
2015	Immunised EP - Artificial immune system inspired mutation and selection	[12]
2016	Mixed mutation strategy EP - Gaussian, Cauchy and Lévy mutations are used	[13]
2017	Fast Convergence EP - An algorithm, which boosts convergence speed and solution quality	[14]
2017	Immune log-normal EP - log-normal mutation combined with artificial immune system	[15]
2018	ADM-EP - automatically designed mutation operators	[16]

See also

• Artificial intelligence
• Genetic algorithm
• Genetic operator