linopy
Linear and mixed-integer optimization with N-D labeled arrays.
PyPSA/linopy Linear and mixed-integer optimization with N-D labeled arrays. linopy is the modeling layer underneath flixopt and fluxopt, so a lot of my contributions started as features I needed for my own work. The rest is open source pulling me in — this is the kind of foundational tooling I want to see well-maintained.
It’s a Python library for building linear and mixed-integer programs using xarray-style labeled N-D arrays. Variables, constraints, and objectives are expressed as array operations rather than scalar-by-scalar loops, which keeps large models compact and fast to construct. It plugs into multiple solver backends (HiGHS, Gurobi, CPLEX, and others), so you’re never locked in.
It lives under PyPSA and is used widely across the energy modeling community. I’m a maintainer.