offlinerlviaio
Science Score: 54.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
✓CITATION.cff file
Found CITATION.cff file -
✓codemeta.json file
Found codemeta.json file -
✓.zenodo.json file
Found .zenodo.json file -
○DOI references
-
✓Academic publication links
Links to: arxiv.org -
○Academic email domains
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (14.3%) to scientific vocabulary
Repository
Basic Info
- Host: GitHub
- Owner: TolgaOk
- License: mit
- Language: Python
- Default Branch: master
- Size: 22.3 MB
Statistics
- Stars: 3
- Watchers: 1
- Forks: 0
- Open Issues: 0
- Releases: 2
Metadata Files
README.md
Offline Reinforcement Learning via Inverse Optimization
This repository provides the source code of the experiments and implementation of the algorithms proposed in the paper.
Installation
To run the provided examples you will need to install MOSEK along with the MOSEK license. MOSEK provides free academic license.
Once the MOSEK installation is completed you can install the required packages and the research package.
bash
pip install -r requirements.txt
pip install -e .
Alternatively, you can use apptainer to build a self contained image using the image.def file. Run start.sh --build to build a apptainer image and run start.sh --run to start a container running vs-code server.
Additional packages
This repository contains several experiments that contains comparison between IO agent and several other RL algorithms. These experiments are run on Quadrotor environment provided in safe-control-gym and MuJoCo control benchmark. In order to run these experiments, an additional installation process is required.
These steps can be done by following the installation process of the listed repositories below.
- safe-control-gym for the Quadrotor environment.
- D4RL for offline MuJoCo control benchmark datasets.
- OfflineRL-Kit for the offline RL agent.
- Stable Baselines3 for the PPO agent.
- For the iterative IO agent:
Examples
You can find the examples under the examples folder:
examples/quadrotor.ipynb: experiments of Sections 4
The experiment directory contains jupyter-notebooks for the corresponding experiments. You can visualize the results within the notebooks.
Citing
Please cite the following work if you found it useful.
bibtex
@misc{dimanidis2025offlinereinforcementlearninginverse,
title={Offline Reinforcement Learning via Inverse Optimization},
author={Ioannis Dimanidis and Tolga Ok and Peyman Mohajerin Esfahani},
year={2025},
eprint={2502.20030},
archivePrefix={arXiv},
primaryClass={cs.LG},
url={https://arxiv.org/abs/2502.20030},
}
Owner
- Name: Tolga Ok
- Login: TolgaOk
- Kind: user
- Website: tolgaok.github.io
- Repositories: 2
- Profile: https://github.com/TolgaOk
Citation (CITATION.cff)
cff-version: 1.2.0
message: "Official Implementation"
authors:
- family-names: Ok
given-names: Tolga
orcid: "https://orcid.org/0000-0002-3669-6121"
title: "Offline Reinforcement Learning via Inverse Optimization"
version: 0.1.0
doi: 10.5281/zenodo.10961728
date-released: 2024-04-11
url: "https://github.com/TolgaOk/offlineRLviaIO"
GitHub Events
Total
- Push event: 1
Last Year
- Push event: 1