Skip to content

Latest commit

 

History

History
33 lines (27 loc) · 1.72 KB

README.md

File metadata and controls

33 lines (27 loc) · 1.72 KB
Raw

Budget Pacing Simulation Framework

Running the Simulation

Running a single simulation

  1. Set up the desired configuration in src/configuration.py.
    1. Choose the desired pacing algorithm by defining the pacing_algorithm variable, e.g.: pacing_algorithm = constants.BudgetPacingAlgorithms.MYSTIQUE_NON_LINEAR.
  2. Run src/main.py.

Running multiple simulations simultaneously

If you're running the simulations through PyCharm, you can use the run configurations defined in the runConfigurations directory. First, set up the desired configuration in src/configuration.py, without changing the value of the pacing_algorithm variable. Then run the runConfigurations/All-Pacing-Systems.run.xml compound configuration, which will run the simulation with each of the following pacing systems:

  1. Mystique Linear
  2. Mystique Non-Linear
  3. Mystique Linear Hard Throttling
  4. Mystique Non-Linear Hard Throttling
  5. Without Budget Pacing

The predefined run configurations set the PACING_SYSTEM environment variable which is read by the src/configuration.py file.

Output

The simulation outputs the statistics into a directory, that is dedicated to the current configuration, which will be created inside the output folder.

Reference

If you use this simulation in one of your research projects, please cite our WWW paper "Mystique: A Budget Pacing System for Performance Optimization in Online Advertising":

@inproceedings{stram2024Mystique,
  title={Mystique: A Budget Pacing System for Performance Optimization in Online Advertising},
  author={Stram, Rotem and Abboud, Rani and Shtoff, Alex and Somekh, Oren and Raviv, Ariel and Koren, Yair},
  booktitle={Proceedings of the ACM Web Conference 2024},
  year={2024}
}