Model-independent observational constraints with fast radio bursts

In this study, we use well-localized fast radio bursts (FRBs) to constrain cosmological parameters through two model-independent approaches: the reconstruction of the Hubble parameter $H(z)$ with an artificial neural network and cosmography. By integrating FRB data with supernovae (SNe), BAO from DESI DR2, and cosmic chronometers (CC), we derive constraints on the Hubble constant (H₀), the deceleration parameter (q₀), and the jerk parameter (j₀). For the reconstruction method, our MCMC analysis with FRB-only provides H₀ = 69.9 ± 5.8 \, km \, s⁻¹ \, Mpc⁻¹, corresponding to a precision of ~ 8%. A joint analysis with FRB+SNe+(BAO+BBN+CMB) gives H₀ = 68.85-0.48+0.47 \, km \, s⁻¹ \, Mpc⁻¹, reaching a precision below $1%$. The cosmographic approach with FRBs alone provides H₀ = 65.83-4.87+3.77 \, km \, s⁻¹ \, Mpc⁻¹, q₀ = -0.45-0.31+0.26, and j₀ = 1.17-0.58+0.70 with a precision for the Hubble constant of ~ 6%. In addition, the BAO+BBN+CMB dataset yields H₀ = 65.20-1.28+1.29 \, km \, s⁻¹ \, Mpc⁻¹, q₀ = -0.29± 0.07, and j₀ = 0.58-0.04+0.03, indicating a precision of ~ 2% for the Hubble constant. Combining the FRB, SNe, BAO+BBN+CMB, and CC data sets provides tighter constraints, for example, H₀ = 67.88-0.53+0.52 \, km \, s⁻¹ \, Mpc⁻¹, q₀ = -0.42-0.03+0.02, and j₀ = 0.56 ± 0.02. In particular, these findings provide a statistically significant indication of deviation from the $Λ$CDM prediction of j₀ = 1, suggesting possible departures from standard cosmology at a confidence level $1σ$. They also confirm a model-independent accelerated expansion (q₀ < 0), demonstrating the potential of FRBs as valuable cosmological probes.