Bonjour,
Posté pour éviter de détourner un fil ailleurs.
Non. En gros, s'il faut un in(dé)fini dans un calcul, ça va. S'il en faut plusieurs, c'est de l'hypercomputing: i.e. au delà des capacités d'une MT. Si les boucles temporelles sont physiquement possibles, il est àmha possible que ça rende l'univers non Turing-équivalent.
Plus intéressant, l'invariance de Lorentz en version quantique n'est probablement pas Turing équivalent. Cf ce message de David Berenstein sur le blog de Moshe
There are theorems about Lorentz invariant quantum mechanics: all such systems that are not trivial, must have infinite dimensional Hilbert space representations. This means that such a system can not be described by a finite number of q-bits.
(...)
As Moshe said: all experimental evidence is in favor of Lorentz invariance (even at the Planck scale).
If you seriously mix gravity with quantum mechanics, no one is sure what is the right answer is, but whatever it is, it should be compatible with what we observe. So if anyone believes that the Universe is strictly Lorentz invariant, one needs something more than an ordinary quantum computer with finite numbers of q-bits to simulate it.
If it is not, there’s some explaining to do as to why we have not seen violations of Lorentz invariance yet (spontaneous violation of Lorentz invariance due to matter effects still falls under the Lorentz invariant paradigm at high energies).(...)
-----