Implementation of a proximal bundle method (PBM) to solve the ACOPF's dual problem, which is formulated as an unconstrained concave maximisation problem with a partially separable structure induced by the clique decomposition. In particular, we use this PBM as a processing step to polish MOSEK's dual solution.
dualACOPFsolver is implemented in Python3. The required packages are:
- numpy
- scipy
- pandas
- osqp
- chompack
- cvxopt
The ACOPF instances are taken from the library PGLib (https://github.com/power-grid-lib/pglib-opf), which is maintained by the IEEE PES Task Force on Benchmarks for Validation of Emerging Power System Algorithms.
Executing
python3 main.py
will run the numerical experiments presented in the paper "A. Oustry, C. D'Ambrosio, L. Liberti, M. Ruiz, Certified and accurate SDP bounds for the ACOPF problem, XXIIth IEEE Power System Computation Conference, Porto, Portugal, June 2022". Executing
python3 stats.py
produces the full result table.
Researchers affiliated with
(o) LIX CNRS, École polytechnique, Institut Polytechnique de Paris, 91128, Palaiseau, France
(o) École des Ponts, 77455 Marne-La-Vallée
Sponsored by Réseau de transport d’électricité, 92073 La Défense, France