Tesis doctoral de Gema M. Martin Reyes
Evolving complexity and similarity in an artificial life framework based on formal language theory although the process of natural selection described by charles darwin the origin of spedes does not guarantee that organisms will increase in complexity as they evolve, since for certain lineages nave been in this way, a big part of the scientific community defend the fact that the tendency of the complexity has been increasing during the evolution. For that reason, many researchers have modelled evolving artificial ecosystems in order to make a case for or against a trend in the evolution of complexity and study the factors that cause it (in the case there exists such a trend). The main problem of these proposed artificial ecosystems is that their resulted can be questionable since they do not use a rigorous complexity measure. This problem comes from the fact that complexity is a complex concept in itself and presents so many variations that it is only valid in specific situations. In this thesis, a formal framework where the evolution of biological complexity can be studied in an objective way is destined. That objectivity is due to state complexity for regular languages is used and it is a well-known and rigorous complexity measure. Such a framework is composed of a population of cyclic unary regular languages (individuals) that try to adapt to a given environment (that also consists of cyclic unary regular languages) by means of evolutionary computation. The genotypes of the individuate are defined as the cyclic unary deterministic finite automata that recognize them and it is shown how they can be represented as binary words. A similarity measure for cyclic unary regular languages is proposed and it is used as fitness function (i.E., The more similar an individual is to the environment, the more adapted to the environment the individual is). To define the species concept and to analyze the disruptive effect that the usual genetic operations produce when they are applied over the genotypes represented as binary words. Many properties and characterizations of such a framework are presented. A relation between the cyclic unary deterministic finite automata and the primitive words is presented, and it provides a characterization of the minimalist of such automata. By using it, two more appropriate representations for the genotypes are introduced. The first one provides a set of operations that preserve permittivity of words and the second one provides a relation between the primitive words and number theory. In this framework, the evolution shows a tendency of the complexity of the individuals to increase. Moreover, results show a strong correlation between the complexity of the population of automata and the complexity of the environment, and that the predatory behaviour promotes the emergence of more complex individuals. By using the framework proposed in this thesis, a wide variety of ecological experimented could be done in a rigorous way.
Datos académicos de la tesis doctoral «Evolving complexity and similarity an artificial life framework based on formal language theory«
- Título de la tesis: Evolving complexity and similarity an artificial life framework based on formal language theory
- Autor: Gema M. Martin Reyes
- Universidad: Málaga
- Fecha de lectura de la tesis: 02/07/2010
Dirección y tribunal
- Director de la tesis
- Francisco José Vico Vela
- Tribunal
- Presidente del tribunal: josé Muñoz pérez
- blanca Truthe (vocal)
- friedrich Otto (vocal)
- mario de Jesús Pérez jiménez (vocal)