PaVeReduce Problem - Githubissues

FeynCalc / feyncalc

Mathematica package for algebraic calculations in elementary particle physics.

https://feyncalc.github.io

GNU General Public License v3.0

310 stars 87 forks source link

PaVeReduce Problem #29

Closed Teddym0913 closed 6 years ago

Teddym0913 commented 6 years ago

Your Mathematica version

9.0 for Mac OS X x86 (64-bit) (February 13, 2013)
Your FeynCalc version

9.2.0
Did you try to reinstall FeynCalc (stable version) using the automatic installer to make sure that you have the latest bugfixes?

Yes, the FeynCalc is installed using the automatic installer just yesterday.
Please provide a minimal working example that illustrates the problem. Please explain the difference between the current behavior and the expected behavior.
Current behavior:
```
In[]:= A0[m1]
Out[]= m1 B0[0, m1, m1]
```
expected behavior:
```
In[]:= A0[m1]
Out[]= m1 + m1 B0[0, m1, m1]
```
Have not yet checked other function carefully. Maybe also has some problem with B1 or B11.

vsht commented 6 years ago

The result given by FeynCalc, i.e.

-((2 m^2 PaVe[0, {0}, {m^2, m^2}])/(2 - D))

is correct:

$LoadAddOns = {"FeynHelpers"}; << FeynCalc`

A0[m^2] // PaXEvaluate PaVeReduce[A0[m^2], A0ToB0 -> True] // PaXEvaluate

gives

m^2/Epsilon - m^2 (-1 + EulerGamma - Log[ScaleMu^2/(m^2 [Pi])])

in both cases. In case you naively did

PaVeReduce[A0[m^2], A0ToB0 -> True] /. D -> 4

then no wonder you got something nonsensical, since after this replacement the finite part of the integral is not correct anymore.

Teddym0913 commented 6 years ago

Thanks for explanation. I just move from History version to the new one, I will look into that and I fount the FAQ page talking about the this. Thanks!

vsht commented 6 years ago

You can still set $LimitTo4 to True to get the old behavior, but I would rather use new methods to avoid it. With FeynHelpers you can easily obtain the full singularity structure with UV and IR divergences nicely separated for almost any 1-loop integral. In the upcoming FeynCalc 9.3 (i.e. the current development version) there is also a function PaVeUVPart (c.f. https://feyncalc.org/forum/1287.html) that will give you the UV singularity of any scalar 1-loop integral. So for renormalization purposes in MS schemes you would not even need FeynHelpers. I currently see no scenartio where would really be forced to use "$LimitTo4". Especially, since with it one can easily make mistakes and end up with a wrong finite part.