D16 Computational Cosmology: Many-Body Simulations

Topic

At the heart of this model lie cosmological N-body simulations, which integrate the trajectories of a vast number of dark matter particles—and, in hydrodynamic versions, gas particles as well—within a cubic volume featuring periodic boundary conditions and the expanding background of the Standard Model.

Over time, benchmark simulations have scaled up from ten billion particles in a moderate volume to tens of billions in a massive volume—representing the largest hydrodynamic simulation run to date.

The tool enabling a comparison between these simulations and actual observations is a statistical framework shared by both: the power spectrum, the correlation function, the halo mass function, and density profiles—all of which can be measured using the very instruments described elsewhere in this document.

The most remarkable feature of these modern simulations is their multi-scale feedback: within a single simulated volume, they simultaneously model cosmology on scales of thousands of megaparsecs and stellar processes within individual galaxies on scales of just a few kiloparsecs, with feedback from supernovae and active galactic nuclei regulating the very star formation that gave rise to them. Without such feedback, simulations produce an excess of stars and halos with incorrect profiles; conversely, with well-calibrated feedback, the simulated universe faithfully reproduces real-world observations.

Furthermore, comparing these simulations with the latest surveys is already revealing tensions that could point to new physics beyond the Standard Model. Thus, from the N-body cosmological simulations that open this document to the tensions hinting at new physics that conclude it—and encompassing everything from the increasing scale of simulated particles to the statistical framework shared with observational data—computational cosmology allows a single simulated universe, governed by its own internal feedback, to be tested against the real universe, step by step, until the point where the two begin to diverge is identified.

ID:4138