Comparison of dark energy models using late-universe observations

In the framework of general relativity, dark energy was proposed to explain the cosmic acceleration. A pivotal inquiry in cosmology is to determine whether dark energy is the cosmological constant, and if not, the challenge lies in constraining how it evolves with time. In this paper, we utilize the latest observational data to constrain some typical dark energy models, and make a comparison for them according to their capabilities of fitting the current data. Our study is confined to late-universe observations, including the baryon acoustic oscillation, type Ia supernova, cosmic chronometer, and strong gravitational lensing time delay data. We employ the Akaike information criterion (AIC), deviance information criterion (DIC), and Bayesian information criterion (BIC) to assess the worth of models. The AIC and DIC analyses indicate that all dark energy models outperform the $\Lambda$CDM model. However, the BIC analysis leaves room for $\Lambda$CDM due to its heavier penalty on the model complexity. Compared to $\Lambda$CDM, most dark energy models are robustly supported by AIC and DIC while being explicitly disfavored by BIC. The models that are robustly favored by AIC and DIC and not explicitly disfavored by BIC include the $w$CDM, interacting dark energy, and Ricci dark energy models. Furthermore, we observe that an alternative modified gravity model exhibits superior performance when compared with $\Lambda$CDM across all information criteria.