A new strategy for testing cosmology with simulations

Abstract

Structural properties of clusters of galaxies have been routinely used to claim either tension or consistency with the fiducial cosmological hypothesis, known as ΛCDM. However, standard approaches are unable to quantify the preference for one hypothesis over another. We advocate using a ‘weighted’ variant of approximate Bayesian computation (ABC), whereby the parameters of the strong lensing-mass scaling relation, α and β, are treated as the summary statistics. We demonstrate then, for the first time, the procedure for estimating the likelihood for observing α and β under the ΛCDM framework. We employ computer simulations for producing mock samples, and account for variation between samples for modelling the likelihood function. We also consider the effects on the likelihood, and consequential ability to compare competing hypotheses, if only simplistic computer simulations are available.