Документ взят из кэша поисковой машины. Адрес оригинального документа : http://zebu.uoregon.edu/~uochep/talks/talks05/jan_strube.pdf Дата изменения: Mon Aug 29 21:24:25 2005 Дата индексирования: Tue Oct 2 11:26:46 2012 Кодировка: Поисковые слова: п п п п п п п п п п п п п п п п п п п п п

Vertexing Tools Vertexing Tools
Jan Strube University of Oregon


Overview Overview
· · · · · · Introduction The Algorithm ­ ZvTop Implementation in hep.lcd Introducing org.lcsim Present Status Summary and Outlook

2005-8-20

Jan Strube - Snowmass 2005

1


The Algorithm ­ ZvTop The Algorithm ­ ZvTop
D. Jackson, NIM A388:247-253, 1997 D. Jackson, NIM A388:247-253, 1997

· Topological Vertex Finding · Looks for overlap of Track probabilities in the Jet · Resolves ambiguities with a resolution criterion · Tracks are fitted to the point of highest overlap
2005-8-20 Jan Strube - Snowmass 2005 2


ZvTop at work ZvTop at work

Tracks are assigned Gaussian Tubes
2005-8-20

The maximum overlap is calculated
3

Jan Strube - Snowmass 2005


PTmiss

PT-corrected Mass for PT-corrected Mass for Flavor Tagging Flavor Tagging
PTVtx

m

Seed Vertex

PT

=

P
e
-

Vtx



tracks



m

±

+ PTVtx

2

2

+ PTVtx

e

+

IP

· The PT-corrected Mass already by itself gives quite reasonable results · But you can see the contamination of B events in the D sample
Jan Strube - Snowmass 2005 4

2005-8-20


ZvTop in hep.lcd ZvTop in hep.lcd
· ZvTop was successfully used in SLD analyses · The Fortran code was ported to hep.lcd by W. Walkowiak and presented here in 2001 · Has been used successfully in studies to extend the algorithm · The code is available and ready to use · Cannot read LCIO files
2005-8-20 Jan Strube - Snowmass 2005 5


Introducing org.lscim Introducing org.lscim
· Concept-independent framework
­ Call for contributions !!!

· Platform-independent implementation
­ Runs everywhere Java 1.5 runs

· Collaboration-independent persistence
­ LCIO format is the new global standard

· Experience-independent learning curve
­ Pick up a CD and start doing an analysis ­ Additional tutorials may be offered upon request
2005-8-20 Jan Strube - Snowmass 2005 6


Under Construction Under Construction
· Code being ported to org.lcsim
­ Better integration into framework ­ Making the code more modular
· Vertex fitting and vertex finding should be decoupled

· The code is available. I would call it an early alpha.
­ A snapshot is distributed with every org.lcsim release ­ More recent versions available in CVS
2005-8-20 Jan Strube - Snowmass 2005 7


Additional Functionality Additional Functionality

This is an actual screenshot from WIRED4 in JAS3 of a J/Psi->mu mu event. The position is the one found by the algorithm, but the tracks are not fitted. The dimensions are just 1, 2, 3 mm in x, y, z, because functionality not yet implemented. This was implemented early to help with debugging 2005-8-20 Jan Strube - Snowmass 2005
8


Work that still can be done Work that still can be done
· Again, the framework is design-independent ! · Associating the vertex with additional neutrals can and should be a natural extension to the PFA that everybody is working on. · Studies by Oregon (J. S. and J. Early, Victoria, 2004) and Oxford (S. Hillert, LCWS 2004) show that flavor tagging can be greatly improved by trying to associate neutrals to the vertex.

2005-8-20

Jan Strube - Snowmass 2005

9


Summary Summary
· Topological Vertex Finding is an important tool for many analyses. · Proven Algorithm · Existing tools allow studies of how to improve the algorithm
­ Has been shown that an extension is useful and important

· Efforts are underway to make tools available to a broader collaboration
2005-8-20 Jan Strube - Snowmass 2005 10