Fully relativistic, cosmological N-body simulations

[matt8282]

Hi,

as you all know, current cosmological N-body simulations (like the Millenium run) are based on the Newtonian limit. Gravitational fields are supposed to be rather weak and therefore the force between dark matter particles reduces to Newtonian gravity. Other relativistic effects are neglected. The cosmological background expands independently of the non-linear processes inside the simulation box. In connection with backreaction studies it might be very useful to have fully-relativistic cosmological N-body simulations. I know that it is extremely difficult to simulate only two merging black holes with all relativistic effects so N-body codes might be completely out of reach. But still, I am interested in this topic. I am not aware of any work on this field.

Can anybody recommend me any article, paper, book or whatsoever? Any source of information is welcome.

Thank you very much in advance!

Best wishes,

Matt

 

[zhermes]

What aspects do you think would require non-linear treatment on cosmological scales?

 

[matt8282]

Well, I don't really have a strong opinion on that. But there are claims that backreaction effects could mimic accelerated expansion (see Buchert, for instance) and standard N-body simulations suppress these effects by construction. So that's why I think it would be nice to check that in fully relativistic simulations (if possible..).

Again, any references concerning this are very welcome!

 

[zhermes]

The backreaction arguments, and associated large-scale inhomogenous concepts, have never been at all convincing to me---but I certainly don't understand them well enough to have a solid opinion. They are certainly not given much credence in the community.

In any case, these effects are only appreciable on scales comparable to the Hubble volume! Just about all long-duration cosmological structure formation simulations are entirely sub-Gpc scale... I just don't see any GR nonlinearities having any effect in that regime.

 

[Nabeshin]

Why fully relativistic? Surely you really don't need all that extra computation, and any kind simulation would be prohibitively expensive. The next step would presumably just be to include the lowest order corrections to Newtonian gravity, although I suspect these terms are neglected for a good reason ;)