Irene is a Python library that solves a broad class of global optimization problems using algebraic methods. It leverages a generalized version of Lasserre's Relaxation technique, making it theoretically applicable to any optimization problem with a well-defined feasible region. Irene's core idea is to solve generalized truncated moment problems within a framework of commutative real algebras.
Documentation & Resources:
Documentation (architecture, solver, benchmarks): irene.readthedocs.io · Install from PyPI: pip install irene
Software archive: doi.org/10.5281/zenodo.23324902
Software paper under peer review at the Journal of Open Source Software (PR openjournals/joss-papers#8635)
Theoretical foundations: The Truncated Moment Problem for Unital Commutative R-Algebras, J. Operator Theory (2023) · Lower bounds for polynomials using geometric programming, SIAM J. Optim. (2012)
Active development: extensions to constraints defined by algebraic differential equations (Differential SDP).