Résumé : We quantify the evolution of matter and galaxy clustering in cosmological hydrodynamical simulations via correlation and bias functions of matter and galaxies. We use simulations TNG100 and TNG300 with epochs from $z=5$ to $z=0$. We calculate spatial correlation functions of galaxies, $\xi(r)$, for simulated galaxies and dark matter (DM) particles to characterise the evolving cosmic web. We find that bias parameters decrease during the evolution, confirming earlier results. At low and medium luminosities, bias parameters of galaxies, $b_0$, are equal, suggesting that dwarf galaxies reside in the same filamentary web as brighter galaxies. Bias parameters of the lowest luminosity galaxies estimated from CFs are lower relative to CFs of particle density-limited clustered samples of DM. We find that bias parameters $b_0$, estimated from CFs of clustered DM, agree with the expected values from the fraction of particles in the clustered population, $b=1/F_c$. The cosmic web contains filamentary structures of various densities, and fractions of matter in the clustered and the unclustered populations are both less than unity. Thus the CF amplitude of the clustered matter is always higher than for all matter, i.e. bias parameter must be $b>1$. Differences between CFs of galaxies and clustered DM suggest that these functions describe different properties of the cosmic web.