Computational Explorations of Life

Author: Chu, Johan

Year: 1999

Degree: Dissertation (Ph.D.)

Advisor: Cross, Michael Clifford

Committee Members: Cross, Michael Clifford; Adami, Christoph Carl; Peck, Charles W.; Wang, Zhen-Gang

Option: Physics

DOI: 10.7907/vkq7-ng96

Abstract

Artificial Life, the creation and study of man-made systems that exhibit the characteristics of life, is a young and still emerging field. The goals of Artificial Life are two-fold; to gain knowledge about and from biology. The Artificial Life system sanda, which extends upon previous systems tierra and avida, was designed to help in investigations into the statistical nature of evolution. As such, it is a model of the simplest living, evolving organisms. Experiments involving tierra, avida, and sanda were the inspiration for investigations into the causes of apparently scale-free dynamics found in these systems. These investigations lead to identification of a branching process that explains the scale-free dynamics of not only these Artificial Life systems, but also those manifested in the taxon rank-frequency distributions of biology and in the size distributions of avalanches in "self-organized critical" sandpile models. This branching process can quantitatively predict-with no free parameters-the pattern of the observed distributions, including their divergence from a true power law. Further, the branching process gives insight into the universal mechanisms involved in the creation of, and divergence from, scale-free dynamics in these systems, including a definition of order and control parameters reminiscent of those from second-order phase transit ions in statistical physics.

Files