Version 5 (modified by md987, 8 years ago) (diff)


How can I implement a new physics model in MadGraph / MadEvent ?

The "old" way : using USRMOD v1

This is "traditional" way to create a new physics model in MadGraph / MadEvent v4. You should use the USRMOD template directory present in the Models directory. The full procedure to follow is extensively described here.

This method has been successfully used over the past to create several new Physics models. However, it presents two important disadvantages:

  • The base model is always the SM and the possible modifications are restricted (e.g. removing a particle is not possible). This was due to the important structural differences in the original BSM models provided with MadGraph / MadEvent v4.
  • Feynman rules should be computed by hand.

To address these issues, two new methods are now available and should be preferentially used (starting from the version 5.0, old models files, including those produced with the USRMOD v1 will only be passively supported).

The "new" way : using FeynRules and/or USRMOD v2


According to the FeynRules website, "FeynRules is a Mathematica® package that allows the calculation of Feynman rules in momentum space for any QFT physics model. The user needs to provide FeynRules with the minimal information required to describe the new model, contained in the so-called model-file. This information is then used to calculate the set of Feynman rules associated with the Lagrangian. The Feynman rules calculated by the code can then be used to implement the new physics model into other existing tools, such as MC generators. This is done via a set of interfaces which are developed together and maintained by the corresponding MC authors."

All interested users should report directly to the FeynRules website for more information on this convenient tool and on the associated interface for MadGraph / MadEvent.

Thanks to its high capabilities, including a user friendly interface, the automatized model generation for arbitrary QFT physics theories and robust validation procedures, FeynRules and the corresponding interface will be the default tool to generate important MadGraph / MadEvent models starting from version 4.6.


In order to keep the USRMOD approach flexibility with the new models produced by the FeynRules package, a new independent version of the USRMOD scripts, called USRMOD v2, has been released. The "philosophy" remains the same, i.e. starting from a base model, the user add/remove/modify particles, interactions and/or parameters, then run the script and finally edit the relevant coupling expressions.

Compare to the first version, it has the following advantages:

  • Fully compatible with (and only with) all models produced by FeynRules, so the base model can be the SM but also any other model previously produced, including those produced by the

USRMOD v2 itself!

  • No limitation for the possible modifications. The

USRMOD v2 propose to first save a "copy" of the base model and adjust all files to reflects the user changes.

  • Safer thanks to additional internal consistency checks
  • Evolutive object-oriented Python code

Note this approach is really complementary to the full FeynRules method. For full models, including a coherent set of particles/interactions (from the QFT perspective), FeynRules should be used. For limited phenomenological studies of particular effects due to one or a few particles, with/or a restricted set of new interactions, the USRMOD= v2 is a fast and efficient alternative.

The main README file and the beta version (now fully functional) distribution of this code is available in attachment. All questions/remarks should be sent to MichelHerquet.

-- Main.MichelHerquet - 02 Mar 2009

Attachments (2)

Download all attachments as: .zip