lpfics 0.0.1

LPFICS aims to enable to find ONE infeasible constraint set if a model is considered infeasible once resolved

LPFICS Lib to find incompatible constraints set in Linear Problems

LPFICS Idea

Objective: This project aims to enable to find ONE incompatible constraint set if a model is considered infeasible once resolved. Note that it MAY NOT be the ONLY ONE set of constraints making your problem infeasible.

This project is based on PuLP (https://github.com/coin-or/pulp) a LP modeler written in Python. PuLP integrates the solver CBC. This solution works with CBC, and GLPK solvers but does not work with Gurobi (as the output is always NotSolved and not Infeasible if the problem is overconstrainted).

Here is a simplified explaination: If a problem is overconstrainted, the problem will:

• test to solve the problem with the first half of the constraints
  • if there is no solution it continues taking the first half (of the half) of the constraints
  • if there is a solution, it will test the other half
    • if the second half has no solution it will continue dividing the second half set of constraints to test it
    • if the second half has a solution it will rearrange the constraints and test again

References

The methodology (quickXplain) on which the program is based is described, discussed and modeled on the following references:

Junker, Ulrich. 2004. « QUICKXPLAIN: Preferred Explanations and Relaxations for Over-Constrained Problems », In: AAAI-04. San Jose. https://www.aaai .org/Papers/AAAI/2004/AAAI04-027.pdf

Morriet, Lou, Benoit Delinchant, Gilles Debizet, Frédéric Wurtz. 2020. « Algorithme d identification de contraintes incompatibles pour les problemes d optimisation : application à un projet energetique ». In: soumis à IBPSA 2020.

Rodler, Patrick. 2020. « Understanding the QuickXPlain Algorithm: Simple Explanation and Formal Proof ». ArXiv:2001.01835, janvier. http://arxiv .org/abs/2001.01835.

Verger, Guillaume, 2016. Ruby-Cbc [Logiciel]. Ruby. In : Github. Consulte le 15 novembre 2020, https://github.com/gverger/ruby-cbc.

Installation Help

Please install LPFICS Lib with pip using the command prompt and type:

pip install git+https://gricad-gitlab.univ-grenoble-alpes.fr/omegalpes/lpfics.git

Please ask [e-mail] if you are struggling during your installation

Library Installation Requirements

PuLP > 1.6.10

Copyright

This project contains a derivative work based on https://github.com/gverger/ruby-cbc. Project ruby-cbc is released under the MIT license, Copyright 2016 Guillaume Verger. Modifications Copyright 2019 G2Elab / MAGE See the LICENSE file for copyright information.

Main Authors:

Guillaume VERGER

Lou MORRIET

Partners:

Univ. Grenoble Alpes, CNRS, Grenoble INP, G2Elab, CEA, Université Paris-Sud

Acknowledgments:

This work has been partially supported by the CDP Eco-SESA receiving fund from the French National Research Agency in the framework of the "Investissements d’avenir" program (ANR-15-IDEX-02), by the ADEME, the French Agency for Environment and Energy Management, with the RETHINE project and by the Region Auvergne Rhône Alpes with the project OREBE.