Документ взят из кэша поисковой машины. Адрес оригинального документа : http://zebu.uoregon.edu/~uochep/talks/talks05/torlumi01-05.pdf Дата изменения: Tue Jan 11 23:07:51 2005 Дата индексирования: Tue Oct 2 11:49:06 2012 Кодировка: ISO8859-5 Поисковые слова: m 43

The Lumi Spectrum Challenge

January 8th, 2005 MDI at the ILC Workshop SLAC

Eric Torrence University of Oregon

Eric Torrence

1/8

January 2005


Energy Measurements

0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 490 492 494 496 498 500 502 504 Root(s) (GeV)

s
+e+e- Еe+e- e e Е
Eric Torrence

e+e- Е ЕЕ
2/8

1 2

1

2 e+e- Еe+e-

January 2005


Luminosity Spectrum "Challenge"

Goals З Demonstrate that we can extract a physics quantity (mH, mt) using only experimental observables З Understand the technical issues in extracting and using the dL / d s information in a physics analysis З Expand interest in this subject worldwide З Better coordinate existing efforts

Caveats З We should concentrate first on something simple as a learning experience but ensure an outcome З This can also be used as leverage to improve the global tools for simulation and analysis З Full-blown "mock data challenge" is not really a good idea at this time
Eric Torrence 3/8 January 2005


Event Generators Precision Measurements З Need precision tools З Must be integral part of MC generators http://home.cern.ch/jadach Don't we know how to do this? Isn't this trivial? I hope so, but I doubt it...

BSL Phase ISR

ISR

Hard Proc

Е Е Need ( p + + p - ) (4 parameters) to generate hard process

Must at least provide s, z distributions (E+, E-) extracted from real data (not just s ) BSL production not independent for each beam
Eric Torrence 4/8 January 2005


Simple Version Generation Layer Simulation Layer Analysis Layer Machine Parameters GuineaPig Physics Generator beam params output beams four-vectors

dL/dE Extraction

Physics

З We should have several different data samples representing different operating conditions З Machine Parameters should be varied within a single data set and should include non-optimal mean values (energy imbalance, jitter, beam offsets, ...) З Glen's IP Feedback files can be put to use here Detail in simulation layer can be increased as time goes on. Probably start with simple parameterized smearing (or even start with no simulation at all?)
Eric Torrence 5/8 January 2005


More Sophistication dL/dE extraction З Ultimately, we use MC as an analysis tool З Simulation of dL/dE must be input directly Extracted dL/dE params dL/dE Analysis "Data" "MC" Physics Analysis dL/dE sim Generator Detector Sim

mH, mtop

Need iterative "bootstrap" procedure to get dL/dE parameters for MC simulation (from Jadach...)

Eric Torrence

6/8

January 2005


Other Thoughts

З With full generation information, events can be reweighted to simulate different machine conditions З Should save four-vectors in format suitable for feeding to full detector simulation some day З Want to leverage existing sophisticated analyses and analyzers (like Klaus Desh) З Try to build framework and documentation to make it easy for casual users to pick this up З Many questions about which approximations are OK and which are not З Important to identify and reserve critical reference reactions (say e+e- Е ) that other detector groups might also want to use for calibration...

Eric Torrence

7/8

January 2005


Proposal I personally think we should get started on something like this soon as an international effort with something to report by Snowmass Immediate To-Do list
Е Е З Characterize p + , p - variations in Glen's files (GP does not write out all information...)

З Investigate technical problems with generation layer З Look at existing formats (LCIO) for defining events or decide to use our own simple format З Evaluate how sophisticated we want to make the simulation layer (start with none?) З Find people to be responsible for analysis layer tasks I see this as a first step in a long program. Should design things to be modular, easily upgradable, and open to international participation (ie: don't tie to closely to one particular detector or analysis package).

Eric Torrence

8/8

January 2005