Performance of various event generators in describing multijet final states at the LHC

TL;DR

This paper evaluates how well different Monte Carlo event generators simulate multijet final states at the LHC by comparing their predictions with ATLAS data, highlighting areas of agreement and discrepancy.

Contribution

It provides a comprehensive comparison of various state-of-the-art event generators against experimental data, identifying specific modeling improvements needed.

Findings

Some generators agree well with data for certain variables.

Discrepancies highlight the need for improved modeling of multijet processes.

The study offers guidance for refining phenomenological models.

Abstract

At the Large Hadron Collider (LHC), the most abundant processes which take place in proton-proton collisions are the generation of multijet events. These final states rely heavily on phenomenological models and perturbative corrections which are not fully understood, and yet for many physics searches at the LHC, multijet processes are an important background to deal with. It is therefore imperative that the modelling of multijet processes is better understood and improved. For this reason, a study has been done with several state-of-the-art Monte Carlo event generators, and their predictions are tested against ATLAS data using the Rivet framework. The results display a mix of agreement and disagreement between the predictions and data, depending on which variables are studied. Several points for improvement on the modelling of multijet processes are stated and discussed.