Could computer simulations lead to discovering new laws of the universe?

[bassplayer142]

Of all our learning of the natural world we have come up with many universal constants. How things interact and such. Whos to say we couldn't make a computer program with these laws in a space and let it go. Disregarding the fact that the memory and computation speed of the computer would be enormous. But wouldn't you be able to learn from this and play around with it until we discover new things out there. And if the program didn't end up like our universe then we know we would be missing many parts.

 

[marcus]

Look at this
http://www.physics.lsa.umich.edu/hubble-volume/expert.htm
http://www.physics.lsa.umich.edu/hubble-volume/
==quote==
* Each simulation:
o employs one billion mass elements and 1024^3 Fourier grid cells
o generates nearly 0.5 terabytes of raw output (later compressed to about 200 Gb)
o requires roughly 70 hours of CPU on 512 processors (four years of a single processor!)

* Some details of the LCDM model :
o Wm = 0.3, WL=0.7, s8 = 0.9, power spectrum from CMBFAST
o simulated cube of comoving length 3/h gigaparsecs (3000/h Mpc)
o simulation begun at redshift z = 35
o force resolution is 0.1/h Mpc
==endquote==

Note that the formidable SIMON WHITE is involved (Cambridge, UC Berkeley, now director Max Planck Astrophysics at Garching)
kick-ass astrophysicist IMHO.

But it is just a simulation with mass points corresponding to galaxies.
It is called Hubble Volume. roughly a cube chunk of the universe on a scale of 13 billion lightyears on a side at present (earlier smaller).

I think that uses effective largescale modeling.

If you are interested in microscopic modeling of spacetime on a Planck scale using Monte Carlo simulation of a quantum gravity dynamical model, then
Renate Loll did that, with her co-worker Jan Ambjorn at Utrecht Netherlands but her computer resources were teensy compared with what she needed.
http://arxiv.org/abs/hep-th/0509010
The Universe from Scratch
R. Loll, J. Ambjorn, J. Jurkiewicz

Her universes were brief little quantum fluctuation burps. But SHE DID NOT FORCE THEM TO BE EVEN THE RIGHT DIMENSION AND THEY TURNED OUT TO BE 3 + 1 = 4 DIMENSIONAL. That was a triumph, which occurred in 2005. After many years many people being frustrated, she succeeded in having it evolve the right dimensionality of its own accord instead of being told what dimension to be. After all, Nature does this.

Dan Christensen has the use of a Beowulf cluster (supercomputer) at the Uni Western Ontario and he has been doing quantum gravity simulations but so far I think this is way too small to be what you are imagining.

Maybe it isn't possible even in 100 years. I don't know what computer resources it would take to do a really satisfying job of simulating the universe.

If all you want to do is simulate GALAXY FORMATION in a fixed spatially flat standardized spacetime. then I think that may have been done. Wallace might know.

Other people may know of other computer simulations of universe(s).

 

[Wallace]

Disregarding the fact that the memory and computation speed of the computer would be enormous.

"Aye there's the Rub" as the Prince of Denmark once said

Simulations of something always ignore some details of the thing being modeled. The trick is to 'cut the fat' so to speak and include as much of the details that make a big difference and ignore the details that make a small difference. The difficulty is that it is often not clear which details are which!

Consider also that the Universe is in fact a giant computer. It has a bunch of information to being with (where stuff is), and is computing the effect of evolving that forward under a set of physical laws. This is all we do when we do computer simulations. Therefore in order to exactly simulate the Universe we need to make another Universe and set that running! Anything less than this contains less information than the thing we are trying to model (the Universe) and hence will be an incomplete and approximate answer.

Note that I havn't even opened the can'o'worms that is the irreducible randomness of quantum mechanics...

 

[neurocomp2003]

sergey's animations are pretty sweet

http://www.physics.mcmaster.ca/~syam/research.html

Oh and there's a paper from nature in his papers section.

 

[setAI]

for simulations that like the universe itself are based on simple rules like cellular automata the complexity of simulations are exponentiating with computational power- which is why 21st century physics belongs to Computer Science and formal mathematics: within decades even unsophisticated brute force simulations on quantum and classical computers will not only accurately model all physical laws- but actually be universes equivalent to ours-

http://arxiv.org/abs/quant-ph/9904050
http://arxiv.org/abs/quant-ph/0011122
http://arxiv.org/abs/quant-ph/0501135

this of course leads to the Simulation Argument

http://www.simulation-argument.com/simulation.html