Title :
Studying network morphology in common developmental genomes
Author :
Antonakopoulos, K.
Author_Institution :
Dept. of Comput. & Inf. Sci., NTNU, Trondheim, Norway
Abstract :
In most EvoDevo (Evolutionary Developmental) systems, genotypes are developed and evolved towards a structural or computational goal utilizing some kind of computational architecture (i.e., structures made of connected elements that may compute). Exploiting a common genotype to develop and evolve different classes of computational architectures towards a common goal has previously been successfully implemented, through common developmental genomes. In this work, we focus at how common genomes exploit the underlying architectures during development and build structure (network morphology) in phenotypes for different problem instances and architecture sizes. Common developmental genomes showed an ability to exploit the size of the architecture by actively involving a larger number of nodes/cells while managed to maintain a small number of neutral and static parts in the evolved structures.
Keywords :
Boolean functions; cellular automata; Boolean networks; EvoDevo; cellular automata; computational architecture; developmental genomes; evolutionary developmental systems; genotypes; network morphology; Automata; Biological cells; Complexity theory; Computational modeling; Computer architecture; Genomics; Morphology; Developmental genomes; L-systems; boolean networks; cellular automata; morphology;
Conference_Titel :
Systems, Man and Cybernetics (SMC), 2014 IEEE International Conference on
Conference_Location :
San Diego, CA
DOI :
10.1109/SMC.2014.6974291