Repository for Adaptive TS Tutorial

This repository contains the data and code for the advanced tutorial paper: Trajectory-Oriented Optimization via Adaptive Thompson Sampling and Grid Refinement: A Tutorial with the adaptive_ts Package

Please see the following sub-directories which walk through the experiments and analysis in the paper:

• repast-SIR-example: Shows the example repastSIR model and experiments
• metarvm-example: Shows the metaRVM case study

Dependencies

• This repository requires several dependencies, which are detailed below

• Any supporting data for the examples is archived via Zenodo: https://doi.org/10.5281/zenodo.20139906

Python

• We recommend installing in a virtual environment like venv (and using requirements.txt) or conda (and using env.yml) for dependencies.

## One of:
python -m venv adaptive-ts-env
source adaptive-ts-env/bin/activate
pip install -r requirements.txt
## Or:
conda create adaptive-ts-env --file=env.yaml --python=3.12 
conda activate adaptive-ts-env

Pipx

• The repastSIR examples use pipx to install the repastSIR model code in its own environment. The pipx instructions are available at https://pipx.pypa.io/stable/how-to/install-pipx/

R

• To run the MetaRVM case study, you will need the following dependencies, which are included in the Docker image (see below)

# install from CRAN
install.packages(c("data.table", "odin", "devtools","pkgbuild","dde","optparse","jsonlite"))
# install from github
devtools::install_github('RESUME-EPI/MetaRVM@main',force=T,upgrade = "never")

Docker

To run the metaRVM case study in an isolated environment, one can use Docker. Docker install instructions are available here: https://docs.docker.com/desktop/

And to pull the docker image to run the experiments:

docker pull dogara/run-metarvm:tutorial

Contact

For questions, please feel free to open an issue or contact:

David O'Gara
Decision and Infrastructure Sciences | Modeling, Simulation, and High-Performance Computing
Argonne National Laboratory
dogara@anl.gov