Missing transverse momentum in ATLAS: current and future performance

During the Run-I data taking period, ATLAS has developed and refined several approaches for measuring missing transverse momentum in proton-proton collisions. Standard calorimeter-based $\mathrm{E}_\mathrm{T}^\mathrm{miss}$ reconstruction techniques have been improved to obtain new levels of precision, while new track-based $\mathrm{p}_\mathrm{T}^\mathrm{miss}$ methods provide for a way to have a second independent measurement of the momentum lost due to particles which do not leave tracks in the inner detectors. While both procedures are individually useful, preliminary studies have shown that combining information from both techniques leads to an improved understanding of missing transverse momentum. Data taking conditions during Run-I varied extensively, especially with respect to the amount of pileup activity present in each event, which provides unique challenges to calorimeter-based $\mathrm{E}_\mathrm{T}^\mathrm{miss}$. Multiple solutions have been demonstrated, including methods which exploit both calorimeter and tracking information. These new techniques have been shown to be both very precise at reconstructing events and stable under high pileup conditions. Run-II is fast approaching, in which high-pileup conditions will become normal. With this in mind, ATLAS has investigated the performance of current $\mathrm{E}_\mathrm{T}^\mathrm{miss}$ methods in simulations of such future events. $\mathrm{E}_\mathrm{T}^\mathrm{miss}$ algorithms are seen to scale well to more extreme conditions than are expected, with efforts ongoing to improve this even further.