A novel method for reconstruction of collision vertices at high-intensity hadron colliders

This paper presents a novel method for the reconstruction of interaction vertices in particle collision data. The algorithm is an agglomerative clustering technique designed for high-luminosity environments in current and future proton-proton colliders. Studies are presented in benchmark scenarios reproducing the LHC data-taking conditions and showing the performance as a function of the number of interactions per bunch crossing. Particular focus will be given to the foreseen data-taking conditions at the Run 3 of the LHC. The proposed algorithm is found to significantly improve the position resolution on the hard-scatter interaction by limiting the contamination of tracks from additional interactions in the vertex fit.