Skip to main content
. 2012 Nov 21;7(11):e48862. doi: 10.1371/journal.pone.0048862

Table 2. Some evolutionary algorithms: G-algorithms.

ALGORITHM COMMENTS REFERENCES
Niching by genotypic fitness sharing Fitness (reproductive opportunity) of individuals is shared amongst members of the same genotypic niche. Maintains diversity. [130], [149]
Learnable Evolution Model (LEM) Uses classifiers to learn the genotypic basis of fitness during an evolutionary run; the inferred basis is used to alter selection to favour those with high predicted fitness, and disfavour those inferred to be deleterious. [131], [164]
Metamodel-assisted EAs A regression model relating fitness to genotype is learned during evolution. The model is used to filter offspring individuals before they are evaluated (if their predicted fitness is low). [78][81]
Efficient Global Optimisation (EGO, ParEGO) A regression model of Gaussian process type is used to relate fitness to genotype (based on a sparse initial sampling of individuals). The model is globally searched to find the individual with the best “expected improvement” in fitness. This individual is then evaluated and used to update the model, and the process iterated. [83], [87], [165]