aiqbench analysis
This research applies machine learning techniques to dynamically optimize relaxation parameters in the Alternating Direction Method of Multipliers (ADMM), specifically targeting Model Predictive Control applications. By learning online updates, the method improves average performance on problem classes while maintaining rigorous convergence guarantees for worst-case scenarios. The approach offers a computationally efficient alternative to penalty updates in OSQP-like architectures without requiring matrix refactorizations.
Learning Over-Relaxation Policies for ADMM with Convergence Guarantees
arXiv:2604.26932v2 Announce Type: replace-cross Abstract: The Alternating Direction Method of Multipliers (ADMM) is a widely used method for structured convex optimization, and its practical performance depends strongly on the choice of penalty and relaxation parameters. Motivated by settings such as Model Predictive Control (MPC), where one repeatedly solves related optimization problems with fixed structure and changing parameter values, we propose learning online updates of the relaxation parameter to improve average performance on problem classes of interest, while guaranteeing that asymptotic convergence is not compromised for the worst-case realization of such problems. This choice is computationally attractive in the Operator Splitting Quadratic Program (OSQP)-like architectures, since adapting relaxation does not trigger the matrix refactorizations associated with penalty updates. We establish convergence guarantees for ADMM with time-varying penalty and relaxation parameters under mild assumptions, and show on benchmark quadratic programs that the resulting learned policies improve both iteration count and wall-clock time on average over baseline OSQP.
Abstract:The Alternating Direction Method of Multipliers (ADMM) is a widely used method for structured convex optimization, and its practical performance depends strongly on the choice of penalty and relaxation parameters. Motivated by settings such as Model Predictive Control (MPC), where one repeatedly solves related optimization problems with fixed structure and changing parameter values, we propose learning online updates of the relaxation parameter to improve average performance on problem classes of interest, while guaranteeing that asymptotic convergence is not compromised for the worst-case realization of such problems. This choice is computationally attractive in the Operator Splitting Quadratic Program (OSQP)-like architectures, since adapting relaxation does not trigger the matrix refactorizations associated with penalty updates. We establish convergence guarantees for ADMM with time-varying penalty and relaxation parameters under mild assumptions, and show on benchmark quadratic programs that the resulting learned policies improve both iteration count and wall-clock time on average over baseline OSQP.
Optimization and Control (math.OC); Machine Learning (cs.LG)
arXiv:2604.26932 [math.OC]
(or arXiv:2604.26932v2 [math.OC] for this version)
https://doi.org/10.48550/arXiv.2604.26932 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 246 words
Related coverage
Convergence of Monte Carlo Optimistic Policy Iteration: Beyond Uniform State-Action Updates
arXiv · Jun 10, 2026
Local and Global Stability in Performative Reinforcement Learning
arXiv · Sep 9, 2026
AI-Assisted Discovery and Construction of a Counterexample to the Convergence of Three-Block ADMM with the Identity Matrix as its Third Constraint Block
arXiv · Aug 17, 2026
Maximum Entropy Inverse Reinforcement Learning for Mean-Field Games with Average Reward
arXiv · Jun 16, 2026