Dwarf galaxies provide us many important clues to understanding of galaxy formation. By using the current version of our own semi-analytic model of galaxy formation, in which cosmic structure forms and evolves based on the cold dark matter model of cosmology, we analyze dwarf galaxies. We find that the model well reproduces many properties such as magnitudes, sizes, and velocity dispersions of, especially, dwarf elliptical galaxies. We also find that the dynamical response of the gravitational potential well of dwarf galaxies to the supernova-induced gas removal plays a very important role to obtain large sizes and small velocity dispersions as observed.

In the last decade observations have been able to probe the evolution of the galaxy luminosity function, in particular showing a variation of its faint-end with redshift. We employ the data of the Cluster-EAGLE project, a set of cosmological, hydrodynamical zoom-in simulations of 30 galaxy clusters, to study the evolution of the galaxy luminostity functions in clusters with redshift. We compile a catalogue of simulated galaxies’ luminosities in the SDSS bands using the E-MILES spectra database, and taking into account dust attenuation. Stacked luminosity functions present little evolution with redshift of the faint-end slope from z=3.5 to z=0, regardless of the cluster mass.

Self-interacting dark matter (SIDM) can create sufficiently large cores in dark matter haloes of dwarf galaxies if the self-interaction cross-section is sufficiently large on scales of dwarf galaxies. Such a large cross-section can be realized without changing the densities and shapes of cluster-size haloes by introducing a velocity dependent cross-section. Lowering the central densities of dwarf-size haloes, however, may change the strength of stellar feedback required to reproduce observed properties of dwarf galaxies such as the luminosity function of the Milky Way’s satellite galaxies. We perform simulations of galaxy formation by employing such a velocity dependent self-interaction cross-section to investigate the coupled effect of SIDM and feedback.