Automation of One-Loop Computations for Scattering Amplitudes and Applications to Collider Phenomenology

This thesis presents the automation of loop computations for the calculation of Next-to-Leading Order contributions to theoretical predictions for particle colliders. We start with the general techniques for performing such predictions and how they can be expressed as perturbative expansions in the coupling constants controlling the strength of particle interactions. This leads to the discussion of the subtleties arising when considering higher order corrections, in particular the methods employed for isolating and canceling the infrared divergences occurring at intermediate steps of the computation. We then introduce the Passarino-Veltman and Ossola-Papadopolous-Pittau loop reduction algorithms. The latter is used by the computer code MadLoop that we wrote specifically for the automation of loop computations. The main part of the thesis focuses on the description of this program, starting with its original loop diagram generation algorithm and how it is embedded within the MadGraph5 environment. An initiation to the usage of the code is given, followed by a discussion of its optimizations where particular attention is paid to the implementation of the open-loop technique. Great details about the validation of MadLoop results against those of other codes are given in appendix B for specific kinematic configurations. We also list in the main text the total rates obtained for various processes. Quantitative information on the runtime speed and numerical stability performances are presented for many processes, each representative of a certain class of complexity. This serves as a comparison benchmark, and shows that realistic studies can be performed for any 2 → 3 and most 2 → 4 processes in the Standard Model. We finish by providing two examples of phenomenology study at the Large Hadron Collider using the tools we developed. The first treats the production of a scalar or pseudo-scalar in association with a top quark pair while the second addresses the tri-boson production channel Z W+ W− with MadSpin simulating the subsequent decay to leptons. We conclude with some insights on MadLoop prospects.