The Basic Study of Artificial Ecosystem Models Using Network-type Assembly-like Language

Abstract

An evolutionary approach to synthetic biology using an artificial ecosystem model, Tierra, has been proposed by T. Ray. In his model, digital creatures, i. e., self-replicating assembly-like programs are coevolving, but this system has been considered not to have the same potential for evolutionary innovation as physical systems. It is considered to be a problem to unlock the full potential of evolution in digital media.
To solve the problem, we have proposed a new model using a network-type assembly-like language, where interaction between creatures is restricted locally like C. Adami’s Avida. In this paper, we have investigated the basic properties of our model, and found (1) the number of creatures with the dominant species continues to change while it does not in Avida, and (2) the former is smaller than the latter under almost the same conditions. These show the possibility that our model has more potential of evolution. We also studied the influence of changing world’s size, mutation rate, and so on.