| R. Calabretta, S. Nolfi, D. Parisi & G. P. Wagner, "Duplication of modules facilitates the evolution of functional specialization", Artificial Life, Vol. 6, pp. 69--84, 2000. |
.... robotics, evolutionary computation is now being used to evolve both the brains and bodies of virtual[19] 1] and real world robots[14] and focus is increasingly coming to bear on making the genetic encoding of these systems as modular and compact as possible in order to increase evolvability[11][4]. Eggenberger[8] first incorporated GRNs into an evolutionary simulation to evolve three dimensional shapes. In this paper I report new results obtained from the Artificial Ontogeny system (AO) which grows virtual agents from GRNs and evaluates them in a physically realistic, three dimensional ....
R. Calabretta, S. Nolfi, D. Parisi & G. P. Wagner, "Duplication of modules facilitates the evolution of functional specialization", Artificial Life, Vol. 6, pp. 69--84, 2000.
.... robotics, evolutionary computation is now being used to evolve both the brains and bodies of virtual[19] 1] and real world robots[14] and focus is increasingly coming to bear on making the genetic encoding of these systems as modular and compact as possible in order to increase evolvability[11][4]. Eggenberger[8] rst incorporated GRNs into an evolutionary simulation to evolve threedimensional shapes. In this paper I report new results obtained from the Arti cial Ontogeny system (AO) which grows virtual agents from GRNs and evaluates them in a physically realistic, three dimensional ....
R. Calabretta, S. Nol , D. Parisi & G. P. Wagner, \Duplication of modules facilitates the evolution of functional specialization", Arti cial Life, Vol. 6, pp. 69-84, 2000.
....than nonmodular ones and they construct a better internal representation of the task. Rueckl et al. 1989) hypothesize that this might be one of the reasons for the evolutionary emergence of the two distinct neural pathways in real organisms. In order to test this hypothesis, Di Ferdinando, Calabretta and Parisi (2000) repeated the experiment of Rueckl et al. 1989) by also allowing the evolution of the network architecture. In a first set of simulations, they used a genetic algorithm for evolving both the architecture and the connection weights of the neural networks. The results showed the inability of the ....
.... that a change affecting a gene will produce a positive effect is reduced with increased pleiotropy of that gene (i.e. by the number of phenotypical characters affected by that gene) A good mapping therefore should reduce pleiotropic effects among characters serving different functions (Calabretta et al. 2000b, p. 77) This kind of efficient mapping can be called modular. More generally, a modular structure of the genotype to phenotype mapping can increase the propensity to evolve (evolvability) of the system possessing this modular structure. However, one should add two considerations. The first ....
[Article contains additional citation context not shown here]
Calabretta, R., Nolfi, S., Parisi, D. & Wagner, G. P. (2000b). Duplication of modules facilitates the evolution of functional specialization. Artificial Life 6:69-84.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC