Kaleu: a general-purpose parton-level phase space generator

TL;DR

The paper tackles the challenge of efficient phase-space generation for multi-particle final states in collider physics. It introduces Kaleu, an independent recursive phase-space generator that uses a vertex-weighted multi-channel approach and adapts on the fly to the integrand, implemented in Fortran for parallel runs. It demonstrates integration with established LO frameworks (e.g., Helac/Phegas, Alpgen) and benchmarks Kaleu across $e^+e^-$ and hadron-collider scenarios, showing competitive performance and optimized phase-space sampling. Overall, Kaleu provides a flexible, parallelizable tool for efficient Monte Carlo integration in parton-level calculations.

Abstract

Kaleu is an independent, true phase space generator. After providing it with some information about the field theory and the particular multi-particle scattering process under consideration, it returns importance sampled random phase space points. Providing it also with the total weight of each generated phase space point, it further adapts to the integration problem on the fly. It is written in Fortran, such that it can independently deal with several scattering processes in parallel.