Timezone: »

Accelerating Quadratic Optimization with Reinforcement Learning
Jeffrey Ichnowski · Paras Jain · Bartolomeo Stellato · Goran Banjac · Michael Luo · Francesco Borrelli · Joseph Gonzalez · Ion Stoica · Ken Goldberg

Fri Dec 10 08:30 AM -- 10:00 AM (PST) @

First-order methods for quadratic optimization such as OSQP are widely used for large-scale machine learning and embedded optimal control, where many related problems must be rapidly solved. These methods face two persistent challenges: manual hyperparameter tuning and convergence time to high-accuracy solutions. To address these, we explore how Reinforcement Learning (RL) can learn a policy to tune parameters to accelerate convergence. In experiments with well-known QP benchmarks we find that our RL policy, RLQP, significantly outperforms state-of-the-art QP solvers by up to 3x. RLQP generalizes surprisingly well to previously unseen problems with varying dimension and structure from different applications, including the QPLIB, Netlib LP and Maros-M{\'e}sz{\'a}ros problems. Code, models, and videos are available at https://berkeleyautomation.github.io/rlqp/.

Author Information

Jeffrey Ichnowski (University of California Berkeley)
Paras Jain (University of California Berkeley)
Bartolomeo Stellato (Massachusetts Institute of Technology)
Goran Banjac (Swiss Federal Institute of Technology)
Michael Luo (University of California Berkeley)
Francesco Borrelli (University of California Berkeley)
Joseph Gonzalez (UC Berkeley)
Ion Stoica (University of California-Berkeley)
Ken Goldberg (UC Berkeley)

More from the Same Authors