Exploring Structure Formation with Cosmological Simulations

[Silviu]

Hello! i am an undergraduate and I started working with a professor doing numerical simulation of cold dark matter. I understand (more or less) the physics behind and the results are very close to observations, which is a good support for the existence of dark matter. However, I can't see the real purpose of cosmological simulations in general (as I am thinking whether should I do this for graduate school or not). Like, yes they reproduce large scale structure and important properties (velocity curve, Tully-Fisher relation, etc.), but other than reproducing observations, what can you actually do. The thing is that a "particle" of dark matter has several millions of solar masses (due to computation limitation) so you definitely can't make predictions about the type of particle we should look for (like mass, whether or not they interact weakly, etc.), for example. So what is the real goal of a physicist doing cosmological simulations? Thank you!

 

[Drakkith]

However, I can't see the real purpose of cosmological simulations in general (as I am thinking whether should I do this for graduate school or not). Like, yes they reproduce large scale structure and important properties (velocity curve, Tully-Fisher relation, etc.), but other than reproducing observations, what can you actually do. The thing is that a "particle" of dark matter has several millions of solar masses (due to computation limitation) so you definitely can't make predictions about the type of particle we should look for (like mass, whether or not they interact weakly, etc.), for example. So what is the real goal of a physicist doing cosmological simulations?

I'm not a cosmologist, but I would think that these simulations allow scientists to verify that certain models and theories are accurate. And this is a very important part of science. We have to be able to make suitable, accurate models in order to verify that the rules of the underlying theory are correct. If you make a simulation that reproduces results compatible with our observations then that tells you something very different than if your results are wildly different from observations.

Also, I'm not sure I agree that you can't make predictions about the different properties of dark matter. Surely your simulation has some kind of parameter for your dark matter particles that relates, in some way, to the real universe, doesn't it?

 

[Silviu]

I'm not a cosmologist, but I would think that these simulations allow scientists to verify that certain models and theories are accurate. And this is a very important part of science. We have to be able to make suitable, accurate models in order to verify that the rules of the underlying theory are correct. If you make a simulation that reproduces results compatible with our observations then that tells you something very different than if your results are wildly different from observations.

Also, I'm not sure I agree that you can't make predictions about the different properties of dark matter. Surely your simulation has some kind of parameter for your dark matter particles that relates, in some way, to the real universe, doesn't it?

My point was, you simulate large scale so you can get properties at that level. And the model is more or less fixed for few tens of years, lately the simulation just increased the accuracy due to the computer powers but I am not sure what prediction can one make using simulations at this level of accuracy.

 

[kimbyd]

Hello! i am an undergraduate and I started working with a professor doing numerical simulation of cold dark matter. I understand (more or less) the physics behind and the results are very close to observations, which is a good support for the existence of dark matter. However, I can't see the real purpose of cosmological simulations in general (as I am thinking whether should I do this for graduate school or not). Like, yes they reproduce large scale structure and important properties (velocity curve, Tully-Fisher relation, etc.), but other than reproducing observations, what can you actually do. The thing is that a "particle" of dark matter has several millions of solar masses (due to computation limitation) so you definitely can't make predictions about the type of particle we should look for (like mass, whether or not they interact weakly, etc.), for example. So what is the real goal of a physicist doing cosmological simulations? Thank you!

I've seen these sorts of simulations used most frequently in terms of understanding structure formation. The precise details of how compact structures like galaxies and galaxy clusters form can only be tackled with simulations.

This sort of thing can potentially be useful in terms of using counts of galaxies with certain properties as a means to rule out certain models. For example, if a simulation predicts almost no large galaxies beyond, say, a redshift of ten, but we observe thousands of them, then that means that simulation has a flaw. Ideally, if the simulation is done well, the discrepancy can be tied to a different physical model of the universe.