How to generate a subset of diagrams

Hi Will,

For those we arrive on this topic first, please read the topic mention by Will, it contains a lot of important warning about the following syntax.

I think that the following syntax is exactly what you ask for.

generate u u > u d vm mu+ / a u u~ d d~ QCD=0

Cheers,

Olivier

On Apr 12, 2013, at 4:31 PM, Will Flanagan <email address hidden> wrote:

> New question #226583 on MadGraph5:
> https://answers.launchpad.net/madgraph5/+question/226583
>
> Hi Madgraph Experts,
>
> This question somewhat relates to my nice discussion with Olivier:
> https://answers.launchpad.net/madgraph5/+question/226332
>
> I am trying to study the kinematics of WBF, but I am having trouble figuring out how to produce only the diagrams of interest. I understand that interference is important, but I don't yet want to consider other diagrams.
>
> In particular if I generate the following diagrams:
> mg5>generate u u > u d vm mu+ / a QCD=0 ; display diagrams
> which produces:
> http://hepr8.physics.tamu.edu/will/uu_udvmmu_no_a.pdf
>
> I would like to produce only the diagrams with WBF production (diagrams 1, 2, 3, 6, 7, and 8) and ignore the others (diagrams 4, 5, 9, 10, 11, and 12). I can get rid of all of these diagrams I don't want by getting rid of all s-channel W bosons ("$$ w+"). Unfortunately I don't want to do this since I do want s-channel W bosons when they are produced by W-Z fusion (diagram 3 and 8).
>
> Are there ways to artificially change the weights of certain diagrams or to generate only a user-selected subset of diagrams? Or is there clever syntax to say "get rid of s-channel W bosons if they come from a quark, but not if they come from W-Z fusion"?
>
> Any suggestions would be greatly appreciated!
> Will
>
> --
> You received this question notification because you are a member of
> MadTeam, which is an answer contact for MadGraph5.

Thanks Olivier Mattelaer, that solved my question.

Thanks once again Olivier!

Actually I should have been able to see that. For some reason, it didn't occur to me that I could just forbid s-channel quarks. You saved me many hours of frustration!

Will

Hi Olivier,

One last question. Is there a way to exclude diagrams that have vertices with four (or more) particles? In particular, I am trying to get rid of diagrams such as this one:

mg5>define ud = u d
mg5>import model mssm
mg5>generate p p > ud ud mul- mul+ / h2 h1 svm z d u a QCD=0 QED=4 ; display diagrams

(Or we could discuss in terms of
mg5>import model heft
mg5>generate g g > h1 g g
where the first diagram has a five particle vertex discussed in arXiv:1106:0522v1)

Anyway, thanks for all the help already. Any additional suggestion on this would be great.

Cheers,
Will

Hi Olivier,

I'm reopening this question just in case you (or anyone) has some good insight. Thanks!

Will

Hi Will,

Sorry for the late reply, I was in Holliday.

> One last question. Is there a way to exclude diagrams that have vertices
> with four (or more) particles? In particular, I am trying to get rid of
> diagrams such as this one:

They are three possible ways to do this. Two based on physical principle:
1) If you have a spectrum where the coupling is zero, then you can use a restrict_card in order to remove all the zero interaction of the model. See MG5 paper for more details)
2) If the interactions is linked to an additional coupling order (like QED/QCD, let's say for the example NP [New Physics]) then you can remove the interactions by doing something like
generate g g > h1 g g NP=0

and one which is not based on physical principle:
1) You can edit the model file in order to remove the unwanted interactions.
You just have to comment (put "#" in front of the line) the definition of the interactions in files interactions.py
of the associated models.

In this case, you are the author of this new model and therefore the responsible for the associated physical aspect (including his validation)
A minimal series of check is
1) that MG5 succeeds to import the model
2) that the "check" command pass on a series of processes
3) compare a series of 1->2 and 2>2 with analytical formula.

Cheers,

Olivier

On Apr 18, 2013, at 11:45 AM, Will Flanagan <email address hidden> wrote:

> Question #226583 on MadGraph5 changed:
> https://answers.launchpad.net/madgraph5/+question/226583
>
> Status: Solved => Open
>
> Will Flanagan is still having a problem:
> Hi Olivier,
>
> I'm reopening this question just in case you (or anyone) has some good
> insight. Thanks!
>
> Will
>
> --
> You received this question notification because you are a member of
> MadTeam, which is an answer contact for MadGraph5.

Thanks again, Olivier