building statistical thermodynamics from cellular automata

[Alex]

Hi everyone,

While spd group is not active, I believe a wider community of spr is
interested in discrete methods as well. A paper 'Integer Lattice Gases
at Equilibrium' is available in arXiv:

http://arxiv.org/abs/cond-mat/0512292

It describes deterministic and reversible lattice gas automaton which
provides semiclassical statistical distributions at equilibrium (like
Planck's oscillators). The cellular automata rule reproduces ideas of
continuation of motion without interactions and detailed balancing in
interactions (or in terms of classical mechanics: 1st and 3rd Newton's
Laws).

This approach does allow building statistics on not mechanical bases
where microdynamics is defined, for example, by ideas of Maxwell
electrodynamics. Could we come to "statistical (not stochastic)
electrodynamics" this way and get explanation of Planck's Radiation Law
from reversible microdynamics?

I would appreciate your feedback. (This study has been presented at the
94th Statistical Mechanics Conference at Rutgers University.)

Happy New Year,
Alex

[Bruno]

This is certainly interesting to know about and I've bookmarked it for
future reference.

My interest comes via the philosophy of physics, and in particular the
simulation arguments that have been fashionable recently, with
otherwise sane people such as Frank J. Tipler putting forward scenarios
that have no real value outside the centre pages of New Scientist. I
presented a paper to the ESSSAT Conference in Barcelona in 2005,
proposing a reductio ad absurdum of simulation arguments in general -
it's at http://www.nugae.com/philosophy/simulate.pdf (I cannot see a
suitable arxiv.org category for philosophy of physics).

In this context it is valuable to get away from the question "can we
simulate the exact laws of the existing universe?", since this is
simply a distraction. Having a cellular automaton that can potentially
exhibit sufficiently complex behaviour to have its own "thermodynamics"
and eventually its own "biology" clarifies the philosophical arguments;
on the other hand, if it turned out that no cellular automaton could
generate sufficient complexity then that would itself be a very
interesting result.

[Alex]

Bruno wrote:
> This is certainly interesting to know about and I've bookmarked it for
> future reference.
>
> My interest comes via the philosophy of physics, and in particular the
> simulation arguments that have been fashionable recently, with
> otherwise sane people such as Frank J. Tipler putting forward scenarios
> that have no real value outside the centre pages of New Scientist. I
> presented a paper to the ESSSAT Conference in Barcelona in 2005,
> proposing a reductio ad absurdum of simulation arguments in general -
> it's at http://www.nugae.com/philosophy/simulate.pdf (I cannot see a
> suitable arxiv.org category for philosophy of physics).

In proposed arXiv reorganization there is a holder for phylosophy:
physics.HI (see http://arxiv.org/new/physics.html).

> In this context it is valuable to get away from the question "can we
> simulate the exact laws of the existing universe?", since this is
> simply a distraction. Having a cellular automaton that can potentially
> exhibit sufficiently complex behaviour to have its own "thermodynamics"
> and eventually its own "biology" clarifies the philosophical arguments;
> on the other hand, if it turned out that no cellular automaton could
> generate sufficient complexity then that would itself be a very
> interesting result.