GAR (General Augmented lagrangian Riccati) is a set of algorithms for the fast solving of proximal and constrained LQR-type linear systems.

It contains an API for defining implicit-dynamics, equality-constrained linear-quadratic problems, and a set of proximal solvers for these:

• serial and parallel block-sparse Riccati-based solvers
• a stagewise-dense serial Riccati solver