Monte Carlo tuning for Multiple Parton Interactions from the ATLAS data (MPI@LHC 2016)

The modelling of Minimum Bias (MB) and Underlying Event (UE) is a crucial component in the description of soft QCD processes. They are both described by multi-parton interactions (MPI) models, the result of proton collisions containing more than one partonic interaction due to collective and beam remnant effects. We present recent ATLAS studies aimed at improving the description of MPI models in Monte Carlo generators. Results of tunes of the Pythia8 MPI parameters to recent ATLAS measurements at 7, 8 and 13 TeV are reported. Studies on the effect of tuning in presence of matching with leading and next-to-leading matrix elements are also presented for processes such as jets and Z-boson production.