Gana is an Algebraic Modeling Language (AML) for Multiscale Modeling and Optimization. Generated programs (Prg) can be subclasses of Multiparametric Mixed Integer Linear Programming (mpMILP).

Installation

The current development checkout requires the Rust extension, gana._native. The Python modeling API is unchanged. Use Python 3.12 or 3.13 for the candidate validation matrix; other versions and wheel platforms are not yet verified.

Install an available release and its dependencies in a fresh environment:

python3.13 -m venv .venv
.venv/bin/python -m pip install gana

On Windows, use .venv\Scripts\python instead. Published releases may predate the backend retirement described here. A compatible wheel includes the native extension and needs no Rust compiler at runtime. A source installation, including this checkout, needs Rust/Cargo 1.88 or newer and a platform linker; validation uses Rust 1.88.0:

.venv/bin/python -m pip install .

Leave GANA_BACKEND unset. The old rust-parameters value is accepted as an alias; python and unknown values are rejected. A missing extension produces installation guidance instead of selecting a Python backend. See the installation guide and native implementation notes.

Features

Gana supports:

• mutable elements, allowing constraints/functions to be updated dynamically
• multiparametric as well as mixed integer programming
• formulation in both matrix form and set-based mathematical program

Moreover, Gana uses a simple format to write constraints/functions.

Elements

Programs in Gana are written using element sets, namely:

1. I - index
2. V - variable
3. P - parameter
4. T - parametric variable

Solvers

The list of natively supported solvers are:

1. PPOPT for multiparametric programming (mp)
2. Gurobi for Mixed Integer Programming (MIP) Programs can, however, be exported as a .mps or .lp file and passed to most solvers.

Illustration

Gana is best run in Jupyter as notebook (.ipynb) files, where programs can be visualized as canonical mathematical programs using set-notation.

Additionally, matrices can be exported.

Purpose

Gana was developed to enable certain functionalities in energia (py). Both were developed as PhD projects and have ample room for improvement. So please reach out to me on cacodcar@gmail.com with suggestions and such.