Coupling computations through space

A conceptual framework for models of coupled computations is developed, based on the assumption that the computations are performed by a population of processing agents whose structure is derived from Turing machines. As a fundamental premise, the agents are embedded in a well-defined space which ultimately provides constraints on the individual movements, thus enabling their autonomous behaviour. The framework takes the form of a taxonomy, according to which Turing machine components are shared among the processing agents. Following the presentation of the taxonomy, the 𝒮𝒯𝒜 coupling model is picked out and further developed, its features and advantages being stressed in relation to the other models. It is argued that the key aspect of this model is that its notion of computation can only be made sense of at the very interaction between the agents and their environment. Weak and strong versions of the 𝒮𝒯𝒜 model are then identified, and an implementation of the latter is briefly discussed. This implementation was performed in an artificial-life system that takes the form of a cellular automata-based architecture of autonomous agents, which allows one to explore the clear-cut notion of space provided by cellular automata. The resulting model is then discussed by contrasting its features with those possessed by some well-known systems. No actual computer run derived from a particular setup of the resulting model is shown