SCQPTH: an efficient differentiable splitting method for convex quadratic programming
This work addresses the need for scalable differentiable QP solvers in optimization and machine learning applications, though it is incremental as it builds on existing ADMM and OSQP frameworks.
The authors tackled the problem of efficiently solving large-scale convex quadratic programs (QPs) with 100-1000 decision variables and thousands of constraints by developing SCQPTH, a differentiable first-order splitting method based on ADMM, which achieved a 1x to 10x improvement in computational efficiency compared to existing differentiable QP solvers.
We present SCQPTH: a differentiable first-order splitting method for convex quadratic programs. The SCQPTH framework is based on the alternating direction method of multipliers (ADMM) and the software implementation is motivated by the state-of-the art solver OSQP: an operating splitting solver for convex quadratic programs (QPs). The SCQPTH software is made available as an open-source python package and contains many similar features including efficient reuse of matrix factorizations, infeasibility detection, automatic scaling and parameter selection. The forward pass algorithm performs operator splitting in the dimension of the original problem space and is therefore suitable for large scale QPs with $100-1000$ decision variables and thousands of constraints. Backpropagation is performed by implicit differentiation of the ADMM fixed-point mapping. Experiments demonstrate that for large scale QPs, SCQPTH can provide a $1\times - 10\times$ improvement in computational efficiency in comparison to existing differentiable QP solvers.