A self contained Jupyter notebook used to generate experimental data for research paper "Softplus Penalty Functions for Constrained Optimization" hosted on arXiv.org
See softplus_penalty_fn.pdf .
Warning: takes several hours to run on a 12 core / 24 thread machine. Reduce n_tasks to reduce the number of experiments.
Alternatively, raw data can be downloaded from results.csv.zip
numpy, pandas, and scipy 1.6.2 (to ensure comparable results)