We study the driven-diffusion equation, describing the dynamics of density fluctuations
δρ(x
→
, t)
in powders or traffic flows. We have performed quite detailed numerical simulations of this equation in one dimension, focusing in particular on the scaling behaviour of the correlation function
<δρ(x
→
, t)δρ(0, 0)>
. One of our motivations was to assess the validity of various theoretical approaches, such as Renormalization Group and different self consistent truncation schemes, to these nonlinear dynamical equations. Although all of them are seen to predict correctly the scaling exponents, only one of them (where the non-exponential nature of the relaxation is taken into account) is able to reproduce satisfactorily the value of the numerical prefactors. Several other interesting issues, such as the noise spectrum of the output current, or the statistics of distance between jams (showing a transition between a ‘laminar’ régime for small noise to a ‘jammed’ régime for higher noise) are also investigated.