PyPRS: A Python Software Package for Parallel Ranking and Selection Procedures

PyPRS is a Python software package specifically developed to solve large-scale ranking and selection (R&S) problems in parallel computing environments. The underlying parallel computing framework is Ray. PyPRS incorporates four well-known parallel procedures:

• The Good Selection Procedure (GSP)
• The Knockout-Tournament (KT) Procedure
• The Parallel Adaptive Survivor Selection (PASS) Procedure
• The Fixed-Budget Knockout-Tournament (FBKT) Procedure

Users can also upload custom procedures to test and compare performance against these built-in procedures.

📋 Prerequisites

• Python 3.10 is recommended for optimal compatibility.
• Required packages: ray==2.44.1, numpy, scipy, matplotlib, mrg32k3a_numba. Install them using:

python -m pip install ray==2.44.1 numpy scipy matplotlib mrg32k3a_numba

📦 Installation

python -m pip install PyPRS

🖥️ How to use

To run PyPRS on a single computer, users just need to execute the GUI.py file located in the UserInterface package in a Python environment：

• If the PyPRS is downloaded from the source repository, users should first navigate to the parent folder of PyPRS folder and then execute the python -m PyPRS.UserInterface.GUI command in the terminal or command prompt.
• If users installed PyPRS using pip, users can directly run python -m PyPRS.UserInterface.GUI in the terminal or command prompt.

Once the command is executed, the Graphical User Interface (GUI) will launch. In the GUI, users can:

• select a procedure
• configure input parameters
• upload required files
• run the procedure