Generalized color structures

[Juergen Reuter]

Generalized color structures (sextets, eps_ijk couplings, etc.) should be supported by the colorizer inside O'Mega.

[ohl]

I just discussed this with Christian and we've found the solution: introduce a set of "ghost" particles that couple only to color singlets and are counted just like the normal ghosts. Each closed loop of these ghosts receives a (-1) so that we get 9 - 1 = 8. This can be derived formally by treating the Higgs as an external field and using the gluon propagator in this external field (i.e. with Higgs insertions resummed).

This can be implemented directly in O'Mega, once the color summation is moved to O'Mega.

[ohl]

OK, I have new implemented CS's algorithm in the O'Mega color factors. WHIZARD should use the result or reimplement it in Fortran (see src/omega/src/color.ml).

An algorithm using the "(i,-i)-Gluons" runs into two problems:

• it works for gl gl -> H, but requires more and more patchwork in higher orders
• it becomes homelessly inefficient, because too many new color structures must be tried out.

[Juergen Reuter]

This now basically depends on ticket #20. Ranking it down.

[Christian Speckner]

I take the liberty to bump this and raise the severity again as other vertices like graviton-gluon-gluon also depend on this.

[ohl]

Since #20 is closed, the Hgg part is done and the generalized color structures can be moved to milestone:v2.1

[Juergen Reuter]

As this is TOs primary target right now, I move it (optimistically to 2.0.2 first), as the ticket ordering confused several other WHIZARD authors. Furthermore rank it high.

[Juergen Reuter]

This was bogus from my side, I confused the tickets. This is the second or third on TOs priority list. Going back to P2 for it.

[Juergen Reuter]

Included a lot of infrastructure in the branch branches/color6, but we have to agree on the data structure for the color flows. Either the rather pedestrian way with Double_Lines and Triple_Lines inside the framework of Module Flow, or the General_Flow set-up. I would rather go for the second one, which is much cleaner and fancier.

[Juergen Reuter]

This needs to be a joint effort. And I guess, we better move it to 2.1.0.

[Juergen Reuter]

That will be also (started to be) resolved with #56. So moving it close.

[Juergen Reuter]

Milestone renamed

[Juergen Reuter]

Milestone renamed

[Juergen Reuter]

Transferred to Gitlab issue 322.