Redshift Space Distortions (RSD) are an apparent anisotropy in the distribution of galaxies due to their peculiar motion. These features are imprinted in the correlation function of galaxies, which describes how these structures distribute around each other. RSD can be represented by a distortions parameter $\beta$, which is strictly related to the growth of cosmic structures. For this reason, measurements of RSD can be exploited to give constraints on the cosmological parameters, such us for example the neutrino mass. Neutrinos are neutral subatomic particles that come with three flavours, the electron, the muon and the tau neutrino. Their mass differences can be measured in the oscillation experiments. Information on the absolute scale of neutrino mass can come from cosmology, since neutrinos leave a characteristic imprint on the large scale structure of the universe. The aim of this thesis is to provide constraints on the accuracy with which neutrino mass can be estimated when expoiting measurements of RSD. In particular we want to describe how the error on the neutrino mass estimate depends on three fundamental parameters of a galaxy redshift survey: the density of the catalogue, the bias of the sample considered and the volume observed. In doing this we make use of the BASICC Simulation from which we extract a series of dark matter halo catalogues, characterized by different value of bias, density and volume. This mock data are analysed via a Markov Chain Monte Carlo procedure, in order to estimate the neutrino mass fraction, using the software package CosmoMC, which has been conveniently modified. In this way we are able to extract a fitting formula describing our measurements, which can be used to forecast the precision reachable in future surveys like Euclid, using this kind of observations.

In this thesis, I have investigated the evolution of the high-redshift (z > 3) AGN population by collecting data from some of the major Chandra and XMM-Newton surveys. The final sample (141 sources) is one of the largest selected at z> 3 in the X- rays and it is characterised by a very high redshift completeness (98%). I derived the spectral slopes and obscurations through a spectral anaysis and I assessed the high-z evolution by deriving the luminosity function and the number counts of the sample. The best representation of the AGN evolution is a pure density evolution (PDE) model: the AGN space density is found to decrease by a factor of 10 from z=3 to z=5. I also found that about 50% of AGN are obscured by large column densities (logNH > 23). By comparing these data with those in the Local Universe, I found a positive evolution of the obscured AGN fraction with redshift, especially for luminous (logLx > 44) AGN. I also studied the gas content of z < 1 AGN-hosting galaxies and compared it with that of inactive galaxies. For the first time, I applied to AGN a method to derive the gas mass previously used for inactive galaxies only. AGN are found to live preferentially in gas-rich galaxies. This result on the one hand can help us in understanding the AGN triggering mechanisms, on the other hand explains why AGN are preferentially hosted by star-forming galaxies.

In this work I present the first measurements of the galaxy stellar mass function (GSMF) from the first public release of the VIPERS catalogue, containing ∼55,000 objects. First, I present the survey design, its scientific goal, the redshift measurements and validation. Then, I provide details about the estimate of galaxy stellar masses, star formation rates, and other physical quantities. I derive the GSMF of different galaxy types (e.g. active and passive galaxies) and as a function of the environment (defined through the local galaxy density contrast). These estimates represent new observational evidence useful to characterise the mechanism of galaxy evolution.

The formation and evolution of galaxy bulges is a greatly debated topic in modern astrophysics. An approach to address this issue is to look at the Galactic bulge, the closest to us. According to some theoretical models, our bulge built-up from the merger of substructures formed from the instability and fragmentation of a proto-disk in the early phases of Galactic evolution. We may have discovered the remnant of one of these substructures: the stellar system Terzan 5. Terzan 5 hosts two stellar populations with different iron abundances, thus suggesting it once was far more massive than today. Moreover, its peculiar chemistry resembles that observed only in the Galactic bulge. In this Thesis we perform a detailed photometric and spectroscopic analysis of this cluster to determine its formation and evolutionary histories. Form the photometric point of view we built a high-resolution differential reddening map in Terzan 5 direction and we measured relative proper motions to separate its member population from the contaminating field stars. This information represents the necessary work to measure the absolute ages of Terzan 5 populations via the Turn-off luminosity method. From the spectroscopic point of view we measured abundances for more than 600 stars belonging to Terzan 5 and its surroundings in order to build the largest field-decontaminated metallicity distribution for this system. We find that the metallicity distribution is extremely wide (more than 1 dex) and we discovered a third, metal-poor and alpha-enhanced population with average [Fe/H]=-0.8. The striking similarity between Terzan 5 and the bulge in terms of their chemical formation and evolution revealed by this Thesis suggests that Terzan 5 formed in situ with the bulge itself. In particular its metal-poor populations trace the early stages of the bulge formation, while its most metal-rich component contains crucial information on the bulge more recent evolution.

Dynamical models of galaxies are a powerful tool to study and understand several astrophysical problems related to galaxy formation and evolution. This thesis is focussed on a particular type of dynamical models, that are widely used in literature, and are based on the solution of the Jeans equations. By means of a numerical Jeans solver code, developed on purpose and able to build state-of-the-art advanced axisymmetric galaxy models, two of the main currently investigated issues in the field of research of early-type galaxies (ETGs) are addressed. The first topic concerns the hot and X-ray emitting gaseous coronae that surround ETGs. The main goal is to explain why flat and rotating galaxies generally exhibit haloes with lower gas temperatures and luminosities with respect to rounder and velocity dispersion supported systems. The second astrophysical problem addressed concerns instead the stellar initial mass function (IMF) of ETGs. Nowadays, this is a very controversial issue due to a growing number of works on ETGs, based on different and independent techniques, that show evidences of a systematic variation of the IMF normalization as a function of galaxy velocity dispersion or mass. These studies are changing the previous opinion that the IMF of ETGs was the same as that of spiral galaxies, and hence universal throughout the whole large family of galaxies.

The Large Magellanic Cloud (LMC) is widely considered as the first step of the cosmological distance ladder, since it contains many different distance indicators. An accurate determination of the distance to the LMC allows one to calibrate these distance indicators that are then used to measure the distance to far objects. The main goal of this thesis is to study the distance and structure of the LMC, as traced by different distance indicators. For these purposes three types of distance indicators were chosen: Classical Cepheids,``hot'' eclipsing binaries and RR Lyrae stars. These objects belong to different stellar populations tracing, in turn, different sub-structures of the LMC. The RR Lyrae stars (age >10 Gyr) are distributed smoothly and likely trace the halo of the LMC. Classical Cepheids are young objects (age 50-200 Myr), mainly located in the bar and spiral arm of the galaxy, while ``hot'' eclipsing binaries mainly trace the star forming regions of the LMC. Furthermore, we have chosen these distance indicators for our study, since the calibration of their zero-points is based on fundamental geometric methods. The ESA cornerstone mission Gaia, launched on 19 December 2013, will measure trigonometric parallaxes for one billion stars with an accuracy of 20 micro-arcsec at V=15 mag, and 200 micro-arcsec at V=20 mag, thus will allow us to calibrate the zero-points of Classical Cepheids, eclipsing binaries and RR Lyrae stars with an unprecedented precision.

In this Thesis, we study the physical properties and the cosmic evolution of AGN and their host galaxies since z∼3. Our analysis exploits samples of star forming galaxies detected with Herschel at far-IR wavelengths (from 70 up to 500 micron) in different extragalactic surveys, such as COSMOS and the deep GOODS (South and North) fields. The broad-band ancillary data available in COSMOS and the GOODS fields, allows us to implement Herschel and Spitzer photometry with multi-wavelength ancillary data. We perform a multicomponent SED-fitting decomposition to decouple the emission due to star formation from that due to AGN accretion, and to estimate both host-galaxy parameters (such as stellar mass, M* and star formation rate, SFR), and nuclear intrinsic bolometric luminosities. We use the individual estimates of AGN bolometric luminosity obtained through SED-fitting decomposition to reconstruct the redshit evolution of the AGN bolometric luminosity function since z∼3. The resulting trends are used to estimate the overall AGN accretion rate density at different cosmic epochs and to trace the first ever estimate of the AGN accretion history from an IR survey. Later on, we focus our study on the connection between AGN accretion and integrated galaxy properties. We analyse the relationships of AGN accretion with galaxy properties in the SFR-M* plane and at different cosmic epochs. Finally, we infer what is the parameter that best correlates with AGN accretion, comparing our results with previous studies and discussing their physical implications in the context of current scenarios of AGN/galaxy evolution.

In this Thesis, we study the accretion of mass and angular momentum onto the disc of spiral galaxies from a global and a local perspective and comparing theory predictions with several observational data. First, we propose a method to measure the specific mass and radial growth rates of stellar discs, based on their star formation rate density profiles and we apply it to a sample of nearby spiral galaxies. We find a positive radial growth rate for almost all galaxies in our sample. Our galaxies grow in size, on average, at one third of the rate at which they grow in mass. Our results are in agreement with theoretical expectations if known scaling relations of disc galaxies are not evolving with time. We also propose a novel method to reconstruct accretion profiles and the local angular momentum of the accreting material from the observed structural and chemical properties of spiral galaxies. Applied to the Milky Way and to one external galaxy, our analysis indicates that accretion occurs at relatively large radii and has a local deficit of angular momentum with respect to the disc. Finally, we show how structure and kinematics of hot gaseous coronae, which are believed to be the source of mass and angular momentum of massive spiral galaxies, can be reconstructed from their angular momentum and entropy distributions. We find that isothermal models with cosmologically motivated angular momentum distributions are compatible with several independent observational constraints. We also consider more complex baroclinic equilibria: we describe a new parametrization for these states, a new self-similar family of solution and a method for reconstructing structure and kinematics from the joint angular momentum/entropy distribution.

This work considers the reconstruction of strong gravitational lenses from their observed effects on the light distribution of background sources. After reviewing the formalism of gravitational lensing and the most common and relevant lens models, new analytical results on the elliptical power law lens are presented, including new expressions for the deflection, potential, shear and magnification, which naturally lead to a fast numerical scheme for practical calculation. The main part of the thesis investigates lens reconstruction with extended sources by means of the forward reconstruction method, in which the lenses and sources are given by parametric models. The numerical realities of the problem make it necessary to find targeted optimisations for the forward method, in order to make it feasible for general applications to modern, high resolution images. The result of these optimisations is presented in the \textsc{Lensed} algorithm. Subsequently, a number of tests for general forward reconstruction methods are created to decouple the influence of sourced from lens reconstructions, in order to objectively demonstrate the constraining power of the reconstruction. The final chapters on lens reconstruction contain two sample applications of the forward method. One is the analysis of images from a strong lensing survey. Such surveys today contain $\sim 100$ strong lenses, and much larger sample sizes are expected in the future, making it necessary to quickly and reliably analyse catalogues of lenses with a fixed model. The second application deals with the opposite situation of a single observation that is to be confronted with different lens models, where the forward method allows for natural model-building. This is demonstrated using an example reconstruction of the ``Cosmic Horseshoe''. An appendix presents an independent work on the use of weak gravitational lensing to investigate theories of modified gravity which exhibit screening in the non-linear regime of structure formation.

21 cm cosmology opens an observational window to previously unexplored cosmological epochs such as the Epoch of Reionization (EoR), the Cosmic Dawn and the Dark Ages using powerful radio interferometers such as the planned Square Kilometer Array (SKA). Among all the other applications which can potentially improve the understanding of standard cosmology, we study the promising opportunity given by measuring the weak gravitational lensing sourced by 21 cm radiation. We performed this study in two different cosmological epochs, at a typical EoR redshift and successively at a post-EoR redshift. We will show how the lensing signal can be reconstructed using a three dimensional optimal quadratic lensing estimator in Fourier space, using single frequency band or combining multiple frequency band measurements. To this purpose, we implemented a simulation pipeline capable of dealing with issues that can not be treated analytically. Considering the current SKA plans, we studied the performance of the quadratic estimator at typical EoR redshifts, for different survey strategies and comparing two thermal noise models for the SKA-Low array. The simulation we performed takes into account the beam of the telescope and the discreteness of visibility measurements. We found that an SKA-Low interferometer should obtain high-fidelity images of the underlying mass distribution in its phase 1 only if several bands are stacked together, covering a redshift range that goes from z=7 to z=11.5. The SKA-Low phase 2, modeled in order to improve the sensitivity of the instrument by almost an order of magnitude, should be capable of providing images with good quality even when the signal is detected within a single frequency band. Considering also the serious effect that foregrounds could have on this detections, we discussed the limits of these results and also the possibility provided by these models of measuring an accurate lensing power spectrum.