SLIDE 1
Multipurpose Event Generators and ep Physics Simon Pltzer IPPP, - - PowerPoint PPT Presentation
Multipurpose Event Generators and ep Physics Simon Pltzer IPPP, - - PowerPoint PPT Presentation
Multipurpose Event Generators and ep Physics Simon Pltzer IPPP, Department of Physics, Durham University & PPT, School of Physics and Astronomy, University of Manchester & Particle Physics, University of Vienna at the VHEeP Workshop
SLIDE 2
SLIDE 3
Event generators overview
Multipurpose Herwig 7 Pythia 8 Sherpa 2 Hard amplitudes some internal, some internal, general LO internal, general via libraries general via ev. fjles loops via libraries Shower options QTilde, pt ordered, CSShower, Dipoles DIRE, VINCIA DIRE NLO Matching internal automated, external internal automated,
- sub. & mult.
S-MC@NLO NLO Merging yes yes yes Hadronization Cluster String Cluster Specialized (ep context): Ariadne (dipoles), Cascade (CCFM), DIPSY (IS evolution), ...
SLIDE 4
Multipurpose Event Generators: State of the Art
Hard process calculation One loop, many legs. Automated. Parton shower algorithms Hard to claim controllable uncertainty. → Multiple interaction models Eikonal or Interleaved, Difgraction (?) Hadronization models New insights into colour reconnection. Hot stufg Pushing showers to higher orders.
SLIDE 5
The Toolbox
T L |T|^2 (T*L) BLHA Extra Native HepMC LHAPDF PDF LHEF & Hooks Ф S &M Ti.Tj Showers, MPI, Hadronization, Decays σ Merging
SLIDE 6
ep Status
All LHC-age multipurpose event generators can simulate DIS. Just showers can't work beyond inclusive stufg – desperately need a good description of hard jets: Matching & merging mandatory. Thanks to general frameworks and automation, ep simulation is possible including all of the state-of-the-art enhancements. Dedicated ep, eA studies by specialized efgorts, e.g. DIPSY LHC physics sets the priority for multipurpose event generators. ep studies highly demanded to cross check assumptions and new development in all (also LHC relevant) physics domains. Little to present though – also high energy studies mostly done for FCC-hh/ee. Remarks and suggestions later, focus on recent (perturbative) development.
[Gustafson, Lönnblad et al. '07 – ]
SLIDE 7
Parton Shower Variations
Aim at evaluating event generator uncertainties in a global prescription → Need to evaluate uncertainties of building blocks one at a time. → Then pin down cross feed, making minimal assumptions. Start with the perturbative part: Parton showers and matching/merging. On-the fmy reweighting available in all multipurpose event generators. Constrain by demanding controllable uncertainties: → Small/large where showers are expected to be reliable/unreliable. → Consistent between systematically difgerent algorithms.
[Bellm, Nail, Plätzer, Schichtel, Siodmok – Eur.Phys.J. C76 (2016) 665] [Bellm, Plätzer, Richardson, Siodmok, Webster – Phys.Rev. D94 (2016) no.3, 034028] [Mrenna, Skands – Phys.Rev. D94 (2016) no.7, 074005] [Bothmann, Schönherr, Schumann –Eur.Phys.J. C76 (2016) no.11, 590]
SLIDE 8
Logarithmic structure
Look at generic Sudakov exponent: AlphaS running on top, also PDF arguments.
[Bellm, Nail, Plätzer, Schichtel, Siodmok – Eur.Phys.J. C76 (2016) 665]
SLIDE 9
Hard Shower Scales
Resummation needs to be cut ofg at a typical hard scale veto on hard emissions, region → to be fjlled by matching. Resummation properties are heavily infmuenced by the way resummation is being switched ofg. Study scale variations in angular ordered and Dipole showers at a benchmark setting where we observe absolutely comparable resummation properties: Hard veto scales, factorization/renormalization scales in the shower and hard process.
[Bellm, Nail, Plätzer, Schichtel, Siodmok – Eur.Phys.J. C76 (2016) 665]
SLIDE 10
Controllable uncertainties – LO
Choice of the hard veto scale is crucial to reproduce hard process input: typically average transverse momenta of hard objects. Controllable uncertainties can
- nly be established by narrow,
smeared versions of a theta function, confjrming simple LL arguments. We can now check the impact of higher order improvements. Still “qualitative” procedure unless showers get higher order corrections.
[Bellm, Nail, Plätzer, Schichtel, Siodmok – Eur.Phys.J. C76 (2016) 665]
SLIDE 11
NLL
NLO Matching
coupling order “accuracy” exclusivity/resolution “jet bin” logarithmic structure “leading” contribution inclusive cross section difgerential cross section LO LL NLO
SLIDE 12
Showers in a nutshell
Showers have virtual and real emission contributions: Showers preserve the total inclusive cross section: Unitarity. Showers approximate tree level matrix elements: In the collinear limits, and in the soft limit for large number of colours N. .
SLIDE 13
The Matching Condition
.
SLIDE 14
Solving the Matching Condition
Infrared cutofg prevents fjnite weights. Add power correction (IR safe observables!) to fjx divergences.
SLIDE 15
Solving the Matching Condition
Infrared cutofg prevents fjnite weights. Add power correction (IR safe observables!) to fjx divergences. dσmatched = + –
SLIDE 16
Solving the Matching Condition
Infrared cutofg prevents fjnite weights. Add power correction (IR safe observables!) to fjx divergences. dσmatched = + –
SLIDE 17
Solving the Matching Condition
Infrared cutofg prevents fjnite weights. Add power correction (IR safe observables!) to fjx divergences. dσmatched = + –
SLIDE 18
NLO Matching
Highly automated – uncertainties and scale setting currently addressed in detail, but no other ambiguities left. This is default for LHC simulation, including complex VBF processes. DIS input needed for understanding And constraining systematics in matching for VBF processes, which show several interesting features. First implementation of DIS matching inside Herwig within this context.
[Rauch, Plätzer – Eur.Phys.J. C77 (2017) no.5, 293] [D'Errico, Richardson – Eur.Phys.J. C72 (2012) 2042]
SLIDE 19
NLL
(N)LO Multijet Merging
coupling order “accuracy” exclusivity/resolution “jet bin” logarithmic structure “leading” contribution inclusive cross section difgerential cross section LO LL NLO
SLIDE 20
NLL
(N)LO Multijet Merging
coupling order “accuracy” exclusivity/resolution “jet bin” logarithmic structure “leading” contribution inclusive cross section difgerential cross section LO LL NLO
SLIDE 21
NLL
(N)LO Multijet Merging
coupling order “accuracy” exclusivity/resolution “jet bin” logarithmic structure “leading” contribution inclusive cross section difgerential cross section LO LL NLO
SLIDE 22
NLL
(N)LO Multijet Merging
coupling order “accuracy” exclusivity/resolution “jet bin” logarithmic structure “leading” contribution inclusive cross section difgerential cross section LO LL NLO
SLIDE 23
Multijet Merging – How?
SLIDE 24
Multijet Merging – How?
“It's complicated.”
SLIDE 25
Basic idea: replace approximate matrix elements with exact ones, but keep Sudakov factors which regularize divergences.
Motivation: Multiple Shower Emissions
.
SLIDE 26
Cut phase space into matrix element and parton shower populated regions.
LO Merging – Phase Space Considerations
.
SLIDE 27
Cut phase space into matrix element and parton shower populated regions.
LO Merging – Phase Space Considerations
.
SLIDE 28
Cut phase space into matrix element and parton shower populated regions.
LO Merging – Phase Space Considerations
.
SLIDE 29
Cut phase space into matrix element and parton shower populated regions.
LO Merging – Phase Space Considerations
.
SLIDE 30
Cut phase space into matrix element and parton shower populated regions.
LO Merging – Phase Space Considerations
.
SLIDE 31
Multijet merging results
Amazing performance for LHC physics – little known for DIS. Specifjcally raises questions about shower scale setting in matching.
[Herwig 7.1 based on Bellm, Plätzer, Gieseke – arXiv:1705.06700]
2
> = 6 8 2 G eV
2
< Q
- 2
< x > = 7 .6 x 1
C l u ster
2
> = 6 8 2 G eV
2
< Q
- 2
< x > = 7 .6 x 1
S tri n g
2
> = 6 8 2 G eV
2
< Q
- 2
< x > = 7 .6 x 1
2
> = 6 8 2 G eV
2
< Q
- 2
< x > = 7 .6 x 1
H 1 D ata
G eV
*
/ d
* T
1/ N dE
1 2 3 4 5 6
2
> = 2 20 G eV
2
< Q
- 1
< x > = 1 .1 x 1
2
> = 2 20 G eV
2
< Q
- 1
< x > = 1 .1 x 1
2
> = 2 20 G eV
2
< Q
- 1
< x > = 1 .1 x 1
2
> = 2 20 G eV
2
< Q
- 1
< x > = 1 .1 x 1
*
- 1 0
1 2 3 4 5 6
2
> = 2 8 3 G eV
2
< Q
- 2
< x > = 2 .6 x 1
2
> = 2 8 3 G eV
2
< Q
- 2
< x > = 2 .6 x 1
2
> = 2 8 3 G eV
2
< Q
- 2
< x > = 2 .6 x 1
2
> = 2 8 3 G eV
2
< Q
- 2
< x > = 2 .6 x 1 1 2 3 4 5 6
2
> = 6 1 7 G eV
2
< Q
- 2
< x > = 2 .6 x 1
2
> = 6 1 7 G eV
2
< Q
- 2
< x > = 2 .6 x 1
2
> = 6 1 7 G eV
2
< Q
- 2
< x > = 2 .6 x 1
2
> = 6 1 7 G eV
2
< Q
- 2
< x > = 2 .6 x 1
2
> = 2 5 3 G eV
2
< Q
- 2
< x> = 1x1
2
> = 2 5 3 G eV
2
< Q
- 2
< x> = 1x1
2
> = 2 5 3 G eV
2
< Q
- 2
< x> = 1x1
2
> = 2 5 3 G eV
2
< Q
- 2
< x> = 1x1 1 2 3 4 5 6
2
> = 5 1 1 G eV
2
< Q
- 2
< x > = 1 .2 x 1
2
> = 5 1 1 G eV
2
< Q
- 2
< x > = 1 .2 x 1
2
> = 5 1 1 G eV
2
< Q
- 2
< x > = 1 .2 x 1
2
> = 5 1 1 G eV
2
< Q
- 2
< x > = 1 .2 x 1
- 1 0
1 2 3 4 5 6
SHERPA SHERPA SHERPA SHERPA
2
> = 1 7 5 G eV
2
< Q
- 3
< x > = 4 .3 x 1
2
> = 1 7 5 G eV
2
< Q
- 3
< x > = 4 .3 x 1
2
> = 1 7 5 G eV
2
< Q
- 3
< x > = 4 .3 x 1
2
> = 1 7 5 G eV
2
< Q
- 3
< x > = 4 .3 x 1 1 2 3 4 5 6
- 1 0
1 2 3 4 5 6
[LO merging addressed in Carli, Gehrmann, Höche – Eur.Phys.J. C67 (2010) 73-97]
SLIDE 32
Shower improvements
Loads of corners to be addressed … and loads of activity (apologies for any omissions) → Electroweak showers → Collinear evolution at higher orders → uPDFs and CCFM → Subleading-N corrections → Amplitude level evolution All require ep as a playground just as ee and pp.
[Christiansen, Sjöstrand] [Krauss, Petrov, Schönherr, Spannowsky] [Bauer, Webber] [Prestel, Höche + Krauss] [Hautmann, Jung] [Plätzer, Sjödahl – n emissions] [Höche, Krauss, Schönherr, Siegert – 1 emission] [Nagy, Soper] [Plätzer]
SLIDE 33
HZTOOL Rivet →
Started to interface HZTOOL to Rivet to push ep studies. Use cases for Cascade, Herwig 7, Pythia 8 but by far not all analysis working out of the box.
[Jung, Plätzer + Prestel – still ongoing and unpublished] DIRE + Pythia 8 [plot by Stefan Prestel] – Cascade [plot by Hannes Jung]
SLIDE 34
Remarks (personal)
Will push on to provide a comprehensive interface from HZTOOL to Rivet. → Will make detailed ep studies with multipurpose event generators much more likely, as threshold of work signifjcantly lowered. → Problem: Several analyses are high maintenance – unfortunately not a one-ofg having done the basic interface. If there are ongoing/planned analyses of HERA data, please use Rivet and do not make model-dependent corrections/assumptions. It would be very nice to see LHC-age multipurpose event generators used in (future) ep collider studies, this will benefjt both communities. An efgort towards a comprehensive (high energy) ep study would be welcome along with a detailed comparison to HERA data.
SLIDE 35