Orateur
Description
The accelerated expansion of the Universe is usually described by a cosmological constant, $\Lambda$. An interesting alternative idea is entropic cosmology, where the entropy associated with the cosmic horizon effectively contributes to (or mimics) dark energy. In this talk, I will compare two commonly discussed routes: (i) holographic dark energy (HDE), where the dark-energy density is tied to a horizon length scale, and (ii) gravity-thermodynamics (GT) models, where the expansion equations are modified by applying the first law of thermodynamics at the horizon.
Using late-time observations of Type Ia supernovae (Pantheon+ and DESy5) together with baryon acoustic oscillation measurements from DESI, we carry out Bayesian parameter estimation and model comparison for generalized entropy deformations motivated by non-standard (Barrow/Tsallis-type) horizon entropies. We find that HDE provides an excellent fit and is statistically competitive with $\Lambda$CDM across data combinations, with an effective equation of state consistent with a mild, quintessence-like evolution.
In contrast, the GT implementations are strongly disfavoured by Bayesian evidence for essentially all data combinations. We then extend the GT framework to broader three- and four-parameter generalized entropy families (which include several popular non-extensive entropies as special limits) and update the analysis with the latest DESI release. The data prefer the standard Bekenstein-Hawking area law, yielding a late-time expansion that is close to $\Lambda$CDM. We also find that the three-parameter entropy description is sufficient and is preferred over the four-parameter extension.
