Skip to content

Software Platform for solving and manipulating multiparametric programs in Python

License

Notifications You must be signed in to change notification settings

TAMUparametric/PPOPT

Repository files navigation

PPOPT

Python package Documentation Status PyPI Downloads Codacy Badge

Python Parametric OPtimization Toolbox (PPOPT) is a software platform for solving and manipulating multiparametric programs in Python.

Installation

Currently, PPOPT requires Python 3.8 or higher and can be installed with the following commands.

pip install ppopt

To install PPOPT and install all optional solvers the following installation is recommended.

pip install ppopt[optional]

In Python 3.11 and beyond there is currently an error with the quadprog package. An alternate version that fixed this error can be installed here.

pip install git+https://github.com/HKaras/quadprog/

Completed Features

  • Solver interface for mpLPs and mpQPs with the following algorithms
    1. Serial and Parallel Combinatorial Algorithms
    2. Serial and Parallel Geometrical Algorithms
    3. Serial and Parallel Graph based Algorithms
  • Solver interface for mpMILPs and mpMIQPs with the following algorithms
    1. Enumeration based algorithm
  • Multiparametric solution export to C++, JavaScript, and Python
  • Plotting utilities
  • Presolver and Conditioning for Multiparametric Programs

Key Applications

  • Explicit Model Predictive Control
  • Multilevel Optimization
  • Integrated Design, Control, and Scheduling
  • Robust Optimization

For more information about Multiparametric programming and it's applications, this paper is a good jumping point.

Quick Overview

To give a fast primer of what we are doing, we are solving multiparametric programming problems (fast) by writing parallel algorithms efficiently. Here is a quick scaling analysis on a large multiparametric program with the combinatorial algorithm.

image image

Here is a benchmark against the state of the art multiparametric programming solvers. All tests run on the Terra Supercomputer at Texas A&M University. Matlab 2021b was used for solvers written in matlab and Python 3.8 was used for PPOPT.

image

Citation

Since a lot of time and effort has gone into PPOPT's development, please cite the following publication if you are using PPOPT for your own research.

@incollection{kenefake2022ppopt,
  title={PPOPT-Multiparametric Solver for Explicit MPC},
  author={Kenefake, Dustin and Pistikopoulos, Efstratios N},
  booktitle={Computer Aided Chemical Engineering},
  volume={51},
  pages={1273--1278},
  year={2022},
  publisher={Elsevier}
}