First published July 1 2016
Generalized Stochastic simulation algorithm for Artificial Chemistry
rtificial chemistries (AC) are useful tools and a simple shortcut for the study of artificial life. In many works, ACs are quite straightforward or simplistic or highly unrealistic (or all combined) but in several works AC are extremely complex. Among them, we focus of Hutton Artificial Chemistry HuAC where reactions act on the nodes of a graph (so-called the atoms) where the connected components composed the actual molecules of the environment. The main works from Hutton are based on a 2D simulator (squirm) with auto-replication and several other properties. This paper proposes a computation framework and software that cancel the need for 2d space simulation in the HuAC while keeping a lot of the features of this chemistry. It relies on the Stochastic Simulation Algorithm that has been here adapted to work on graph structure. In order to test it, we simulated Huttons auto-replication which relies heavily on strong spatial inter- actions in a spaceless environment. In addition, due to the increase in performance, we develop some preliminary work on Random Chemical Worlds where reactions are randomly selected. We showed on simple metrics that the fraction of reactions among all possible is a general parameter that acts on the system similarly to a phase transition.