Vertex reconstruction in CMS

Due to the high track multiplicity in the final states expected in proton collisions at the LHC experiments, novel vertex reconstruction algorithms are required. The vertex reconstruction problem can be decomposed into a pattern recognition problem (“vertex finding”) and an estimation problem (“vertex fitting”). Starting from least-squares methods, robustifications of the classical algorithms are discussed and the statistical properties of the novel methods are shown. A whole set of different approaches for the vertex finding problem is presented and compared in relevant physics channels.