MadEvent: Automatic Event Generation with MadGraph

TL;DR

MadEvent automates high-multiplicity event generation by coupling MadGraph’s automatic amplitude and subprocess handling with a new Single-Diagram-Enhanced multi-channel integration to efficiently integrate over complex phase spaces. The core idea is to decompose the squared amplitude into a sum of channel-specific contributions tied to individual Feynman diagrams, enabling independent, parallelizable integrations and targeted sampling. The paper demonstrates the approach on collider-relevant processes (W/Z+jets, ttbar+jets, WW/ZZ, ttbarH, etc.), yielding cross sections and unweighted events that agree with established codes where compared, and providing a public path to showering programs. Limitations and future work include scaling to very large diagram counts and pursuing amplitude factorization and color techniques to extend applicability to more QCD partons. The work offers a practical, automated framework for precise background modeling in current and future collider studies.

Abstract

We present a new multi-channel integration method and its implementation in the multi-purpose event generator MadEvent, which is based on MadGraph. Given a process, MadGraph automatically identifies all the relevant subprocesses, generates both the amplitudes and the mappings needed for an efficient integration over the phase space, and passes them to MadEvent. As a result, a process-specific, stand-alone code is produced that allows the user to calculate cross sections and produce unweighted events in a standard output format. Several examples are given for processes that are relevant for physics studies at present and forthcoming colliders.