ndandanov / pyomo_simple_minlp

A repository to demonstrate a simple MINLP in Pyomo.

A repository to demonstrate a simple MINLP in Pyomo

This repository demonstrates a simple mixed-integer nonlinear optimization problem (MINLP), its modeling and solution in Pyomo. The problem is modeled both with a NonNegativeIntegers domain for the elements of the vector variable and with mixed domains of the vector variable (NonNegativeIntegers and NonNegativeReals).

The following files are included:

• MINLP_description.pdf - a mathematical description of the MINLP optimization problem;

• Jupyter Notebooks:

1. clean_opt_integer.ipynb - the optimization problem with a NonNegativeIntegers domain for the elements of the vector variable z; solved with a locally installed Couenne solver, gives identical results as the BARON solver;
2. clean_opt_integer-neos.ipynb - the same optimization problem, but solved with the NEOS server and Couenne solver, slightly different results;
3. clean_opt_mixed.ipynb - the optimization problem with mixed domains (NonNegativeIntegers and NonNegativeReals) for the elements of the vector variable z; solving with a locally installed Couenne solver does not converge;
4. clean_opt_mixed-neos.ipynb - the same mixed optimization problem, but solved with the NEOS server and Couenne solver, different (presumably incorrect) results than the BARON solver.

• NL model files:

1. mymodel-couenne-integer.nl - model file for notebook 1;
2. mymodel-couenne-integer-neos.nl - model file for notebook 2;
3. mymodel-couenne-neos.nl - model file for notebook 4.