Skip to contents

PIQP is an Proximal Interior Point Quadratic Programming solver, which can solve dense and sparse quadratic programs described in described in Schwan, Jiang, Kuhn, and Jones (2023) (https://arxiv.org/abs/2304.00290). Combining an infeasible interior point method with the proximal method of multipliers, the algorithm can handle ill-conditioned convex QP problems without the need for linear independence of the constraints. The solver is written in header only 'C++ 14' leveraging the Eigen library for vectorized linear algebra. For small dense problems, vectorized instructions and cache locality can be exploited more efficiently. Allocation free problem updates and re-solves are also provided.

Author

Balasubramanian Narasimhan, Roland Schwan (C), Yuning Jiang, Daniel Kuhn, Colin N. Jones