Improving primary-vertex reconstruction with a minimum-cost lifted multicut graph partitioning algorithm

Abstract

Particle physics experiments often require the simultaneous reconstruction of many interaction vertices. Usually, this problem is solved by ad hoc heuristic algorithms. We propose a universal approach to address the multiple vertex finding through a principled formulation as a minimum-cost lifted multicut problem. The suggested algorithm is tested in a typical LHC environment with multiple proton-proton interaction vertices. Reconstruction errors caused by the particle detectors complicate the solution and require the introduction of special metrics to assess the vertex-finding performance. We demonstrate that the minimum-cost lifted multicut approach outperforms heuristic algorithms and works well up to the highest vertex multiplicity expected at the LHC.