Warm-starting active-set solvers using graph neural networks

TL;DR

This paper introduces a graph neural network-based method to predict active constraints in quadratic programming solvers, significantly reducing computation time in real-time control applications.

Contribution

The paper presents a novel GNN approach that exploits problem structure to effectively warm-start active-set QP solvers, generalizing across different problem sizes.

Findings

GNN reduces solver iterations compared to cold-starting.

Performance is comparable to baseline neural network models.

Method generalizes to unseen problem dimensions.

Abstract

Quadratic programming (QP) solvers are widely used in real-time control and optimization, but their computational cost often limits applicability in time-critical settings. To resolve this, we propose a learning-to-optimize approach using graph neural networks (GNNs) to predict active constraints in the dual active-set solver DAQP. Our method exploits the structural properties of QPs by representing them as bipartite graphs and learns to approximate the optimal active set for effectively warm-starting the solver. Across varying problem sizes, the GNN consistently reduces the number of solver iterations compared to cold-starting, while performance is comparable to a multilayer perceptron baseline. In contrast to the baseline, our GNN-based approach trained on varying problem sizes generalizes to unseen dimensions, demonstrating flexibility and scalability. These results highlight the potential of structure-aware learning to accelerate optimization in real-time applications such as model predictive control.