The use and optimisation of integration grid techniques to generate next-to-leading order predictions of jet cross-sections, independent of parton distribution functions, was investigated. Such methods were found to provide an accurate approximation to a standard Monte-Carlo simulation (within 1%) and enable collider data to be readily included in global PDF fitting procedures. However, the benefit of including inclusive-jet cross-section data from ATLAS in global fits is only significant if the jet energy scale (JES) can be constrained to ~1% at high pT. Uncertainties in the theoretical prediction of the inclusive-jet cross-section such as PDFs and fixed-order (scale) uncertainties were studied and compared with experimental errors arising from jet energy resolution and absolute scale. These uncertainties were then considered in the context of a quark compositeness search where a sensitivity to a compositeness scale of Lambda10TeV can be achieved with 10 inverse femtobarns of data, if the jet energy scale can be constrained to ~1%. An analysis using dijet angular distributions found a similar sensitivity without the dependence on the jet energy scale. A potential method of evaluating the stability of the jet energy scale out to high pT by Hbootstrapping the calibration at low pT by the use of multi-jet events was also investigated. This suggests that a calorimeter non-linearity can be detected for jets with pT500GeV at ~1.5%/500GeV (i.e. a 1.5% change in JES over 500GeV in pT). An investigation of inner-detector commissioning issues associated with the ATLAS Semiconductor Tracker (SCT), including a review of Hnoisy modules on the SCT Barrel (from May 2007) was carried out. In addition a tool for DCS monitoring within the online monitoring framework was developed and tested during the M5 and M6 commissioning weeks. Finally, a method of assessing the track reconstruction efficiency by track-insertion was considered for the particular case of minimum bias events.