Cosmological Parameter Estimation using Particle Swarm Optimization

TL;DR

This paper introduces Particle Swarm Optimization (PSO) as an efficient alternative to traditional methods for estimating cosmological parameters from observational data, demonstrating its effectiveness and potential to accelerate analysis.

Contribution

It applies PSO algorithms to cosmological data analysis, showing they can find best-fit parameters faster and serve as a complementary tool to MCMC methods.

Findings

PSO successfully estimates cosmological parameters from observational data.

PSO achieves comparable results to MCMC with less computational time.

PSO outputs can enhance the efficiency of MCMC analyses.

Abstract

The search for the model or ingredients that describe the current vision of our cosmos has led to the creation of a set of highly favorable experiments, and therefore a great flow of information. Due to this torrent of information and the need to analyze it exhaustively, the main aim of this paper is to introduce the Particle Swarm Optimization (PSO) as a complement to traditional cosmological data analysis. The PSO is one of the most representative Bio-inspired algorithms as provides excellent robustness in high-dimensional or complex problems with relative simplicity and small number of parameters during the implementation. In this work we implemented two versions of the canonical PSO algorithm: global best and local best, to explore dark energy models in the light of Type Ia Supernovae and Baryonic Acoustic Oscillations observations, in particular, DESI and DESI+Union3 datasets. The results achieved validate the performance of the PSO algorithm in finding the best-fit parameters from observational data and confirm that PSO, under certain conditions, can deliver competitive results, at a fraction of time, compared to standard MCMC methods. Finally, the PSO output can also serve as a valuable input to the MCMC methods to speed up its analysis.