The repository for the paper ProfitPilot: Enabling Rebalancing in PCNs through Profitable Cycle Creation published on IEEE Transactions on Network and Service Management. The repository contains a simple PCN simulator written in Python with multiple attachment strategies and the scripts used to get the results presented in the paper.
This code uses snapshots from the Lightning Network to test different strategies. Currently, part of the code related to the node placement considers an undirected graph and the part of the code related to rebalancing considers a directed graph. As this is an ongoing project, this will be corrected in the future. On the same note, part of the LN snapshot is available in the repository under the directory ln-snapshot. However, the snapshots used in the rebalancing part is not available because of their size. Users may download the rest of the snapshots here or by cloning and running the code present in here. All snapshots were obtained from the lnresearch repository.
The code requires some Python3 libraries as well. They can be installed by running the following command in the root repository directory:
pip install -r requirements.txt
The program also requires two datasets which are sampled to simulate payments. The first one is a credit-card dataset that can be downloaded from Kaggle. The second one is a Ripple dataset available here. Both datasets must be downloaded and moved to a directory called datasets on the root repository directory.
LN-Looprebalance allows multiple simulation scenarios. To view simulation options, run the following command:
python3 get-results.py --help