Skip to main content

Research Repository

Advanced Search

Maximum Individual Complexity is Indefinitely Scalable in Geb

Channon

Maximum Individual Complexity is Indefinitely Scalable in Geb Thumbnail


Authors



Abstract

Geb was the first artificial life system to be classified as exhibiting open-ended evolutionary dynamics according to Bedau and Packard’s evolutionary activity measures and is the only one to have been classified as such according to the enhanced version of that classification scheme. Its evolution is driven by biotic selection, that is (approximately) by natural selection rather than artificial selection. Whether or not Geb can generate an indefinite increase in maximum individual complexity is evaluated here by scaling two parameters: world length (which bounds population size) and the maximum number of neurons per individual. Maximum individual complexity is found to be asymptotically bounded when scaling either parameter alone. However, maximum individual complexity is found to be indefinitely scalable, to the extent evaluated so far (with runtimes in years and billions of reproductions per run), when scaling both world length and the maximum number of neurons per individual, together. Further, maximum individual complexity is shown to scale logarithmically with (the lower of) maximum population size and maximum number of neurons per individual. This raises interesting questions and lines of thought about the feasibility of achieving complex results within open-ended evolutionary systems and how to improve on this order of complexity growth.

Citation

Channon. (2019). Maximum Individual Complexity is Indefinitely Scalable in Geb. Artificial Life, 134-144. https://doi.org/10.1162/artl_a_00285

Acceptance Date Feb 19, 2019
Publication Date May 31, 2019
Journal Artificial Life
Print ISSN 1064-5462
Publisher Massachusetts Institute of Technology Press
Pages 134-144
DOI https://doi.org/10.1162/artl_a_00285
Keywords open-ended evolution, biotic selection, ongoing growth of complexity, diversity, indefinite scalability
Publisher URL https://doi.org/10.1162/artl_a_00285

Files






You might also like



Downloadable Citations