Monte Carlo phase space integration of multiparticle cross sections with carlomat_4.5

Abstract

Multidimensional phase space integrals must be calculated in order to obtain predictions for total or differential cross sections, or to simulate unweighted events of multiparticle reactions. The corresponding matrix elements, already in the leading order, receive contributions typically from dozens of thousands of the Feynman diagrams, many of which often involve strong peaks due to denominators of some Feynman propagators approaching their minima. As the number of peaks exceeds by far the number of integration variables, such integrals can practically be performed within the multichannel Monte Carlo approach, with different phase space parameterizations, each designed to smooth possibly a few peaks at a time. This obviously requires a lot different phase space parameterizations which, if possible, should be generated and combined in a single multichannel Monte Carlo procedure in a fully automatic way. A few different approaches to the calculation of the multidimensional phase space integrals have been incorporated in version 4.5 of the multipurpose Monte Carlo program carlomat. The present work illustrates how carlomat_4.5 can facilitate the challenging task of calculating the multidimensional phase space integrals.