Comparison with OpEn

[drex2500]

Hi, I came across the ProxDDP paper and was wondering if anyone has used OpEn and knows how it compares for constrained optimisation. Both this method and OpEn seem to have similar problem statements and have been used for MPC. I'm wondering, is this library is faster or more robust? Or am I completely off-track and OpEn is a different thing?

[ManifoldFR]

Hello @drex2500 !

Thanks for your interest in aligator.
I wasn't aware of OpEn's existence. It seems it's a solver engine in Rust with a Python/MATLAB Casadi frontend which generates code for the Rust API?

It's quite different from the philosophy of aligator's API, which is OOP-based (this has its drawbacks, and I'm looking to change it) around writing dynamics, cost, constraint, integrator classes and so on. We also have native Lie group support. We don't have the automatic differentiation capabilities that casadi provides however (we'd need Lie group autodiff and some tighter support for caching some of Pinocchio's more expensive calculations, for performance reasons). I don't think anyone has run aligator on embedded hardware, also -- since we don't do code generation, either.