Teaching, Learning, and Applying Optimization: AMPL’s Intuitive Modeling Meets the Python Ecosystem

– A more natural approach to describing optimization problems. Students can write many common logical conditions, “not-quite-linear” functions, and nonlinear functions the way they think about them, without having to learn complicated and error prone reformulations.

– A Python-first alternative to learning AMPL and model building. New teaching materials leverage the power of Jupyter notebooks and Google Colab to bring modern computing to the study of optimization.

– Faster, easier importing of data and exporting of results. The AMPL Python interface (amplpy) efficiently connects model sets and parameters to Python’s native data structures and Pandas dataframes. An all-new spreadsheet interface reads and writes .xlsx and .csv files, with added support for two dimensional spreadsheet tables.

– Streamlined application  development. Python scripts can be turned quickly into illustrative applications using amplpy, Pandas, and the Streamlit app framework.