Data Science Optimization Toolkit

Software Requirement

• Python 3.9
• The optimization toolkit framework is built on OpenAI Gymnasium

Directory Organization

├── env.yaml                          : Conda setup file with package requirements
├── setup.py                          : Python setup file with requirements files
├── README.md                         : Readme documentation
├── utests                            : Folder containing a collection of unit tests
├── jlab_opt_control
    ├── agents                        : Folder containing different agents
    ├── buffers                       : Folder containing different buffers
    ├── cfgs                          : Folder containing configuration filesfor agents and environments
    ├── core                          : Folder containing base classes
    ├── drivers                       : Folder containing workflow modules / drivers
    ├── envs                          : Folder containing different environemments
    ├── models                        : Folder containing different models
    ├── utils                         : Folder containing supporting tools (e.g. monitoring)

Installing

• Clone code from repo and move into directory

git clone https://github.com/JeffersonLab/SciOptControlToolkit.git
cd SciOptControlToolkit

• Create default conda environment setup: (only once)

conda env create --file env.yaml

• Activate conda environment: (required every time you use the package)

conda activate jlab_opt_control_env

• Install the package in environment (only once)

pip install -e .

Documentation

Cite this software

@misc{SOCT,
  author = {Malachi Schram, Kishan Rajput, Armen Kasparian},
  title = {Scientific Optimization Control Toolkit (SOCT)},
  year = {2024},
  publisher = {GitHub},
  journal = {GitHub repository},
  howpublished = {\url{JeffersonLab/SciOptControlToolkit}},
}

Contacts

If you have any questions or concerns regarding SOCT, please contact Malachi Schram (schram@jlab.org), Kishan Rajput (kishan@jlab.org), Armen Kasparian (armenk@jlab.org).