Hierarchical structure formation, where massive objects form from mergers
of smaller ones is a fundamental prediction of the latest most popular model 
of the universe. This universe has a low (Cold Dark) Matter density,  
a flat curvature (due to the presence of a non-zero cosmological constant), 
and is initially homogeneous and isotropic. In such a universe, density contrast 
grows by gravitational amplification of initially slightly overdense regions.
When they collapse, these regions form virialized "halos" within which baryonic 
gas can cool radiatively and form stars: a disc galaxy is born. 
Obviously, these halos (and the galaxies within them) are not isolated objects 
and they can interact/merge with one another.
This makes it difficult to talk about formation and evolution of galaxies as the 
two processes are intertwined: is formation to be defined as the moment when a galaxy 
had formed half of the stars that it now contains?
Is it the instant at which half of the stars were assembled together in a unique 
object for the first time? How is it possible to reproduce the properties of local 
galaxies in such a scenario? What are its predictions for the high redshift population? 
In this talk, I will present a model which was developed to answer such questions 
and which code name is GALICS for "GALaxies In Cosmological Simulations". 
I will first describe the hybrid method on which GALICS is based i.e. the coupling 
between high-resolution N-body CDM simulations and a semi-analytic modelling of the 
baryon component. I will discuss multi-wavelength results obtained with such a method 
for local galaxies and their redshift evolution up to z = 3 and how one can do a "direct" 
comparison of model and data through mock observations.
Finally, I will conclude with more general considerations on how to fake present and 
future observations with GALICS and dicuss possible extension/improvement of the model.