Signal Region Optimisation Studies Based on BDT and Multi-Bin Approaches in the Context of Supersymmetry Searches in Hadronic Final States with the ATLAS Detector

The searches for supersymmetric phenomena are mostly based on simple Cut & Count methods. One example is the search for squarks and gluinos in final states with multiple jets, missing transverse momentum and without leptons. This analysis, based on $36.1\,\text{fb}^{−1}$ of $pp$ collision data at $\sqrt{s}$ = 13 TeV recorded with the ATLAS detector, uses Cut & Count based methods in the signal regions. In order to improve the analysis sensitivity, the use of sophisticated techniques, such as boosted decision trees (BDT) and Multi-Bin, is being investigated in this thesis. The focus of the study lies on squarks and gluino searches. These techniques are evaluated using Monte Carlo simulation. The goal is to find a new approach which is on the one hand simple but allows for a significant improvement. A gain up to approximately 200 GeV in the neutralino mass and an enhancement of about 200 GeV in the squark and gluino mass is achieved with these new techniques.