At the LHC, experimental data from proton collisions are compared to theoretical predictions of the SM. These predictions are often evaluated using Monte Carlo simulations. In order to ensure the long term success of the LHC it is necessary to improve the theoretical accuracy and computational efficiency of these simulations. This is especially pertinent in the era of the High-Luminosity LHC, where it is important to take advantage of the improved precision of experimental data. In this thesis we present three projects that will help towards this goal, all of which focus on the computation of perturbative physics. The first of these projects is the development of a new phase space generator called SingularPhasespace. Rather than attempting to simulate nature, this generator is used to populate the singular regions of phase space in a controlled and predictable way. This will allow developers to quickly understand any unexpected behaviour in their NLO simulations and so make it easier to improve them. The second project is a restructuring of the MadGraph5_aMC@NLO code, designed to improve its computational efficiency. This optimisation focussed on the evaluation of the fixed-order matrix element. We ensure that the program recycles duplicate parts of the helicity amplitude calculation instead of re-evaluating them. This results in a speed-up of a factor of ~2x for complex processes. We have dubbed this optimisation helicity recycling. The third project is a study of Higgs plus jet(s) simulations at NLO. We used Herwig's internal implementation of both subtractive and multiplicative matching methods to combine NLO fixed-order matrix elements with both the angular-ordered and dipole shower. We find good agreement between the showers but disagreement between the matching methods for certain observables. In particular we focus on the azimuthal angle between jets and Higgs pT distributions. We conclude that the multiplicative matching method is giving the more theoretically accurate description.
|Date of Award||31 Dec 2021|
- The University of Manchester
|Supervisor||Michael Seymour (Supervisor) & Mrinal Dasgupta (Supervisor)|
- Monte Carlo