Abstract
Amoeba is a computer model designed to facilitate the study of the origin and evolution of digital life. Specifically, an initially disordered system, consisting of random sequences of machine instructions, self-organizes into an ordered system containing self-replicating programs. The current version of Amoeba broadens the original system's capability by using a basis set of 32 machine instructions that is computationally universal. In addition, Amoeba uses a set of 64 address labels, each of which is randomly assigned to a machine instruction each time a sequence is randomly created. This eliminates the constraint that occurs when the complements of predefined codons are used for addressing. A more open-ended system results because programs can now form subroutines that are arranged in an arbitrary manner.