Overall Detector Performance Working Group

1
Overall Detector PerformanceWorking Group

• Convenors Pascal Gay,
• Markus Schumacher , Mark Thomson

Prague, 17.11. 02

• Charge of the WG
• Strategy
• How to do ?
• Who does what ?
• Open Questions
• Towards Amsterdam

2
Charge of the Working Group

• Evaluation of the detector performance
  considering the whole detector
• Comparison of different sub-detector
  design/technology options in the context of the
  detector as a single entity
• Provide a forum for discussion on all issues
  related to the overall detector performance with
  participation from detector RD, simulation and
  physics groups (including our colleagues from
  the North American and Asian LC-Workshops)

3
Strategy

• Obtain key performance figures
• Perform reconstruction of physics events using
  information from all subdetectors with full
  simulation
• Study influence of machine backgrounds
  (occupancy), overlaping events, time structure of
  accelerator,
  (mis)allignment and calibration issues,
  issue of crossing angle of the
  colliding beams
• Develop ID/reco. tools and compare them using
  well defined benchmark processes
• Extract parametrisations for / transfer
  algorithms to fast simulation packages (e.g. done
  for flavor tagging)
• Perform comparison of physics results between
  full and fast simulation
  different detector technology
  reco. algorithm options
• Provide inputs to the discussion of
  cost/performance issues

4
How to do it ?

• common simulation framework for a fair comprison
  agree on language and one (maybe) new
  package (so far
  Fortran/C, BRAHMS/MOKKA, GEANT3/4)
• provide a common data format for easy use
  in physics studies
• define and agree on benchmark processes for
  comparsion (eegWW,ZZ,ttH for jet reco., jet
  separation, E resolution) (eegWW gqqlnu for low
  angle tracking, flavor charge ID)
• Identify key analysis tools to be developed
  ( reconstruction
  algorithms, ID tools,
  accelerator alignment issues)
  
  
• Identify key performance numbers to be determined
  
  ( E, p, d0 resolutions, reco and tagging
  eff., fake rates ... )
• More detailed and complete list of
  benchmark processes, tools and
  key performance numbers on working group web page

5
Who does what ?

• detector groups provide inputs about
  technology options,
  detector design (granularity, material budget,
  readout time), very basic performance figures
  (point/energy/cluster resolutions, dependence on
  number of cells etc.), noise, ocupancy ....
• simulation/software groups
  implement or ensure
  implementation of above in simulation,
  
  provide common simulation framework and data
  format
• detector groups and simulation/software groups
  develop and
  implement basic reconstruction algorithms
  ( providing tracks, clusters )
• overall detector performance group and
  simulation/software groups provide/ensure
  existence of analysis tools
  ( e.g. event reconstruction,
  ID packages )
• overall detector performance group
  define
  performance criteria
  collect result
  from various groups and distribute those
  information
• overall detector performance group and all
  (including physics groups)
  evaluate the
  performance figures / compare options

6
Analysis tools/topics and Performance Criteria

• Tools and Topics
• PID reconstruction for e,m
  ,p0,V0,h0,g ,conversions, ....
• quark flavor and charge tag
• dE/dx id algorithms
• implementation of overlapping
  events, alignment calibration, several
  bunch crossings, beam crossing angle

• Performance Criteria
• define physics processes to be studied as
  benchmarks
• specify figures to compare (e.g.)
  reco./id/selection effiencies rejection
  factors / fake rates resolutions D E, D Mij ,
  D Mmiss, ...
• compare those figures for 1) different
  event topologies 2) technology
  reco. options 3) background levels, allignment
  and calibration accuracies

7
Some Open Questions and Topics

• influence of machine backgrounds due to and
  time structure of accelerator (vs.
  readout time)
  (e.g.
  assignment of tracks to bunch crossings)
• issue of (mis)alignment, calibration, dead
  channels
• hermeticity, forward veto vs. machine
  backgrounds
• influence of non vanishing crossing angle
• new and/or more sophisticated analysis tools
  (e.g. quark charge tagging, b (c)
  vs. bbar (cbar) sophisticated
  implementation of dE/dx for particle ID)
  

( no complete list, examples from personal
preference)
8
Towards Amsterdam

• Today
  get overview of
  detector technologies, their implementation and
  status of simulation and availability and
  sophistication of ID/analysis tools
• Today Next Weeks
  1) find volunteers ? for
  all the open topics 2) decide
  on and provide common simulation
  framework and data format
• Until Amsterdam
  1) establish common
  framework and data format 2) write
  documentation and provide it to all users
  3) exchange information between working groups
  4) do first studies for evaluation of
  performance
• Amsterdam
  plenty of reports on new
  and interesting results