On countings and enumerations of block-parallel automata networks
share
When we focus on finite dynamical systems from both the computability/complexity and the modelling standpoints, automata networks seem to be a particularly appropriate mathematical model on which theory shall be developed. In this paper, automata networks are finite collections of entities (the automata), each automaton having its own set of possible states, which interact with each other over discrete time, interactions being defined as local functions allowing the automata to change their state according to the states of their neighbourhoods. The studies on this model of computation have underlined the very importance of the way (i.e. the schedule) according to which the automata update their states, namely the update modes which can be deterministic, periodic, fair, or not. Indeed, a given network may admit numerous underlying dynamics, these latter depending highly on the update modes under which we let the former evolve. In this paper, we pay attention to a new kind of deterministic, periodic and fair update mode family introduced recently in a modelling framework, called the block-parallel update modes by duality with the well-known and studied block-sequential update modes. More precisely, in the general context of automata networks, this work aims at presenting what distinguish block-parallel update modes from block-sequential ones, and at counting and enumerating them: in absolute terms, by keeping only representatives leading to distinct dynamics, and by keeping only representatives giving rise to distinct isomorphic limit dynamics. Put together, this paper constitutes a first theoretical analysis of these update modes and their impact on automata networks dynamics.
READ FULL TEXT