How to run

Use conda to create an environment with the necessary dependencies:

conda env create -f environment.yaml -y
conda activate benchmark

Run the script with python run_solver.py <solver> <file_url>. For example:

python run_solver.py highs https://storage.googleapis.com/solver-benchmarks/benchmark-test-model.lp

An outgoing internet connection is required in order to download the necessary model from our public GCP bucket. The default download location is /tmp/{filename}. The script also accepts a third optional parameter with which you can change the default download location:

python run_solver.py highs https://storage.googleapis.com/solver-benchmarks/benchmark-test-model.lp /home/user/model.lp

At the end of a successful run a JSON object should appear on the standard output. Example:

{"runtime": 285.923513174057, "status": "ok", "condition": "optimal", "objective": 2813303766.241783, "duality_gap": null, "max_integrality_violation": null}

The runtime value is the measured runtime duration in seconds.