2015/04/12 by Nachum Dershowitz, Evgenia Falkovich
Computer Science · #cs.LO
paper · pdf · doi:10.4204/eptcs.179.2
published as EPTCS 179, 2015, pp. 17-32 · In Proceedings DCM 2014, arXiv:1504.01927
arxiv created 2015/04/12 · arxiv updated 2015/04/14
Any algorithm (in the sense of Gurevich's abstract-state-machine axiomatization of classical algorithms) operating over any arbitrary unordered domain can be simulated by a dynamic cellular automaton, that is, by a pattern-directed cellular automaton with unconstrained topology and with the power to create new cells. The advantage is that the latter is closer to physical reality. The overhead of our simulation is quadratic.