摘要:
   Galaxies are believed to reside in dark matter halos, and their formation and evolution are closely related to those of the host halos. Empirical models such as halo occupation distribution (HOD) and conditional luminosity function (CLF) are widely utilized to establish connections between galaxies and halos, as well as to model galaxy clustering. Traditional empirical models assume that galaxy properties only depend on halo mass. However, it is found that galaxy properties can also depend on halo assembly history and environment in semi-analytic models and hydrodynamic simulations. I will introduce a modification to the HOD model to account for this secondary dependence, which could be important for constraining galaxy-halo relations and cosmological parameters in observation. Given the intricate and multivariate nature of the galaxy-halo relation, machine learning techniques are also employed to enhance the recovery of this relation in simulations. I will show my work on predicting galaxy properties with halo properties, which could also be used to compare the galaxy-halo relation in simulations and observation.