The search and score-based waterfall optimization
This code aims to optimize waterfall auction strategy. The code is based on Halbersberg et al., 2022 - https://arxiv.org/pdf/2201.06409.pdf The code is inspired by the well known K2 algorithm (Cooper and Herscovits, 1992)
The repository is organized as follows:
classes - all the necessary classes
data - the synthetic datasets and waterfalls
models - the S&S algorithms: run_algorithms.py apply the S&S/MCTS while search_and_score.py contains the ulilizations - e.g., the neighbor selection
Quick start: run main.py