Reactive Turing machines

We propose reactive Turing machines, extending classical Turing machines with a process-theoretical notion of interaction. We show that every effective transition system is simulated up to branching bisimilarity by a reactive Turing machine, and that every computable transition system with a bounded branching degree is simulated up to divergence-preserving branching bisimilarity by a reactive Turing machine. We conclude from these results that there exist universal reactive Turing machines, and that the parallel composition of a finite number of (communicating) reactive Turing machines can be simulated by a single reactive Turing machine. We also establish a correspondence between reactive Turing machines and the process theory TCPtau, proving that a transition system can be simulated, up to branching bisimilarity, by a reactive Turing machine if, and only if, it is definable by a finite TCPtau-specification.