Features

Efficient

OSQP uses a specialized ADMM-based first-order method with custom sparse linear algebra routines that exploit structure in problem data.

Robust

The algorithm is absolutely division free after the setup and it requires no assumptions on problem data (the problem only needs to be convex). It just works!

Free

OSQP is free and will always be free for everyone. Its license is Apache 2.0.

Embeddable

OSQP has an easy interface to generate customized embeddable C code with no memory manager required.

Many interfaces

OSQP supports many interfaces including C/C++, Fortran, Matlab, Python, R, Julia, Rust.

Library-free

OSQP is self-contained and requires no external library to run.


OSQP beats most QP solvers.

We benchmarked OSQP against problems from many different classes, applications and scalings. OSQP beats most available commercial and academic solvers.


MIT
Lyft
Stanford University
Aptiv
UC Berkeley
Google
Toyota
ETH
KU Leuven
BlackRock
UC Davis
Los Alamos
University of Stuttgart
NYU
IIT
Lund University
Linkedin
Macquarie
Seervision
Siemens

Credits

Bartolomeo Stellato
Bartolomeo Stellato

Main developer

Princeton University

Goran Banjac
Goran Banjac

Main developer

Paul Goulart
Paul Goulart

Methods and Maths

University of Oxford

Stephen Boyd
Stephen Boyd

Methods and Maths

Stanford University

Alberto Bemporad
Alberto Bemporad

Methods and Maths

IMT Lucca

Nicholas Moehle
Nicholas Moehle

Methods, Maths and Code Generation

Stanford University

Michel Schubiger
Michel Schubiger

GPU implementation

Schindler R&D

John Lygeros
John Lygeros

Methods and Maths

ETH Zurich

Ian McInerney
Ian McInerney

Software Engineering and Code Generation

Imperial College London

Vineet Bansal
Vineet Bansal

Software Engineering

CSML, Princeton University

And many others...