Contextual Genetic Algorithms: Evolving Developmental Rules

L.M. Rocha
Computer Research Group, MS P990
Los Alamos National Laboratory
Los Alamos, NM 87545

In: Advances in Artificial Life . F. Moran, A. Moreno, J.J. Merelo, and P. Chacon (Eds.). Series: Lecture Notes in Artificial Intelligence, Springer-Verlag. pp. 368-382.

NOTE: The full version of this paper can be dowloaded in "zipped" postscript (.zip) or Adobe Acrobat (.pdf) format. Here only the abstract and introduction are shown due to the extensive use of mathematical symbols and figures. Notice that in some systems you may have to press the "shift" key while clicking on this link.

Abstract. A genetic algorithm scheme with a stochastic genotype/phenotype relation is proposed. The mechanisms responsible for this intermediate level of uncertainty, are inspired by the biological system of RNA editing found in a variety of organisms. In biological systems, RNA editing represents a significant and potentially regulatory step in gene expression. The artificial algorithm here presented, will propose the evolution of such regulatory steps as an aid to the modeling of differentiated development of artificial organisms according to environmental, contextual, constraints. This mechanism of genetic string editing will then be utilized in the definition of a genetic algorithm scheme, with good scaling and evolutionary properties, in which phenotypes are represented by mathematical structures based on fuzzy set and evidence theories.

1. Introduction

    The essence of GA's lies on the separation of the description of a solution (e.g. a machine) from the solution itself: variation is applied solely to the descriptions, while the respective solutions are evaluated, and the whole selected according to this evaluation [14]. A genetic algorithm "is primarily concerned with producing variants having a high probability of success in the environment" [19, page 35]. Nonetheless, one important difference between evolutionary computation and biological genetic systems, lies precisely on the connection between descriptions and solutions, between signifier and signified. In genetic algorithms the relation between the two is linear and direct: one description, one solution. While in the biological genetic systems there exists a multitude of processes, taking place between the transcription of a description and its expression, responsible for the establishment of an uncertain relation between signifier and signified, that is, a one-to-many relation.

     "The proteins encoded by [DNA] are [...] oxymorphic: their individual shapes are precisely unpredictable. So long as this is true, the genomic language, like our own languages, will not have a logical link between signifier and signified. This will not prevent its being read or understood; rather, it will assure that DNA remains a language expressing as full a range of meanings through arbitrary signifiers as any other language." [26, p. 70]

    In other words, the same genotype will not always produce the same phenotype; rather, many phenotypes can be produced by one genotype depending on changes in the environmental context. If the effects of changing environmental contexts affecting gene expression within an individual can be harnessed and used to it's selective advantage in a changing environment, then we can say that such an individual has achieved a degree of control over its own genetic expression. It is the objective of this paper to propose a computational scheme which may be able to achieve this degree of control. It will be further suggested, that the modeling of biological development may be linked precisely to GA's capable of evolving this extra degree of control.

    To establish this one-to-many relationship between descriptions and solutions in GA's, I will propose an extra mechanism inspired by the edition of RNA in biological genetic systems. Section 2 will introduce some of the known mechanisms of RNA Editing. Section 3 will introduce semiotic model offering a theoretical framework for RNA editing. Section 4 will propose computational counterparts to RNA editing. Section 5 will discuss the utilization of such mechanisms regarding the problem of development. Finally, Section 6 will present one particular algorithm, which utilizes fuzzy set and evidence theory to introduce even higher levels of uncertainty to description/solution relations.