Vertex reconstruction

The code available in this repository implements vertex reconstruction for simulated events in a vertex detector, evaluating its efficiency and resolution.

Running the reconstruction

To compile the reconstruction, it is sufficient to run ROOT and execute the following command:

.x Compile_Reconstruction.cpp+

To run the reconstruction execute the analysis function:

RunAnalysis()

It is possible to add some arguments to the Compile_Reconstruction.cpp file to activate different options:

• .x Compile_Reconstruction.cpp+("force") to force the compilation of the different files
• .x Compile_Reconstruction.cpp+("clean") to delete all the compilation outputs coming from previous compilations before compiling

Configuration of the reconstruction

It is possible to configure a reconstruction using a .yaml file, following these steps:

• Open the .yaml file in a text editor or configuration tool.
• Review the available parameters and their descriptions.
• Modify the values of the parameters as desired to configure the reconstruction.
• Save the .yaml file once you have finished making your changes.

The config file can then be passed to the reconstruction function: RunAnalysis("Config_file.yaml")

Here is an example of settings used to run a reconstruction.

A description of the settings is available here.

Program implementation

Here some details on how the program implements the key steps to implement vertex reconstruction as well as efficiency and resolution evaluation:

• Tracklet reconstruction: we consider tracklets all pairs of intersections with first and second detector whose angle difference $ \varphi = \lvert \varphi_1-\varphi_2\rvert < \varphi_{max} $ with $ \varphi_{max} $ to be setted in the configuration file.

• Tracklet vertex reconstruction: for each reconstructed tracklet, vertex is reconstructed as the intersection between the track and a plane orthogonal to the track in the $xy$ plane passing through the $ z $ axis.

• Event vertex reconstruction: for each event, $ z $ vertex coordinates of all tracklets are collected in an histogram. Then, the program considers the range of values defined by lower and upper edge of the bin with highest frequency: the event vertex is estimated with the running window technique. This ensures combinatorial tracklets do not contribute to the estimated value.

• Setup performance evaluation: after processing all events of the run, the program saves the following plots in an output file

  • Residuals
  • Efficiency vs Multiplicity
  • Resolution vs Multiplicity
  • Efficiency vs true coordinate $ z $ of vertexes
  • Resolution vs true coordinate $ z $ of vertexes