The halo of the MilkyWay (MW) contains residual evidence of its hierarchical accretion history, such as stellar streams, dwarf satellite galaxies and possibly even intermediate-mass black holes the latter carried as they fell into the larger Galaxy. The discovery and study of these objects have the potential to answer elusive questions about our Galaxy, such as the accurate determination of its total mass, a fundamental quantity that determines the properties and fate of galaxies in the Universe.

I use a particle tagging technique to dynamically populate the N-body Via Lactea II high-resolution simulation with stars. The method is calibrated using the observed luminosity function of Milky Way satellites and the concentration of their stellar populations, and self-consistently follows the accretion and disruption of progenitor dwarfs and the build-up of the stellar halo in a cosmological "live host". Simple prescriptions for assigning stellar populations to collisionless particles are able to reproduce many properties of the observed Milky Way halo and its surviving dwarf satellites, like velocity dispersions, sizes, brightness profiles, metallicities, and spatial distribution. I apply a standard mass estimation algorithm based on Jeans modelling of the line-of-sight velocity dispersion profiles to the simulated dwarf spheroidals, and test the accuracy of this technique. The inner mass-luminosity relation for currently detectable satellites is nearly flat in this model, in qualitative agreement with the "common mass scale" found in Milky Way dwarfs.

I extend the tagging approach to the study of intermediate-mass black holes (IMBHs), and assess the size, properties, and detectability of the leftover accreted halo population. The method assigns a black hole to the most tightly bound central particle of each subhalo at infall according to an extrapolation of the MBH-sigma star relation, and self-consistently follows the accretion and disruption of Milky Way progenitor dwarfs and their holes in a cosmological "live" host from high redshift to today. I show that, depending on the minimum stellar velocity dispersion, below which central black holes are assumed to be increasingly rare, as many as two thousand or as few as seventy IMBHs may be left wandering in the halo of the Milky Way today. I identify two main Galactic subpopulations, "naked" IMBHs, whose host subhalos were totally destroyed after infall, and "clothed" IMBHs residing in dark matter satellites that survived tidal stripping. Naked IMBHs typically constitute about half of the total and are more centrally concentrated. Their detection may provide an observational tool to constrain the formation history of massive

black holes in the early Universe.

I use the results from the stellar halo tagging in combination with the state-of-the-art hydrodynamical cosmological simulation Eris to address the question of the poorly known Milky Way halo mass. Taking advantage of the two simulated galaxies' very different masses, I explore the full range of estimates for the Galaxy from observational data. I establish that the simulated halos reproduce many of the properties of the MW stellar halo, including its density profile slope, velocity anisotropy and, in the case of the lighter galaxy, its radial velocity dispersion profile. There is a striking link between discontinuities in these quantities where significant pileup of stars in the orbital apocenters of their progenitors exists in phase space. I carry out controlled experiments using numerical integration of the Jeans equation to conclude that the lighter halo, Eris, indeed provides a much better fit to the data than the more massive halo of Via Lactea II.