Learning-Augmented Algorithms for Online Subset Sum

This is the code for the paper "Learning-Augmented Algorithms for Online Subset Sum"

Requirements

Python >= 3.6.11

Pytorch >= 1.7.0

Adversary Train

To train the adversary network for each algorithm and output the bad instance set, run

bash train_adversary.sh

After training, the instance file for each algorithm will be created.

Algorithm Test

To test the performance of each algorithm in its corresponding instance file, run

bash test.sh