Neural Solvers for Fast and Accurate Numerical Optimal Control
Abstract
Synthesizing optimal controllers for dynamical systems often involves solving optimization problems with hard real-time constraints. These constraints determine the class of numerical methods that can be applied: computationally expensive but accurate numerical routines are replaced by fast and inaccurate methods, trading inference time for solution accuracy. This paper provides techniques to improve the quality of optimized control policies given a fixed computational budget. We achieve the above via a hypersolvers approach, which hybridizes a differential equation solver and a neural network. The performance is evaluated in direct and receding-horizon optimal control tasks in both low and high dimensions, where the proposed approach shows consistent Pareto improvements in solution accuracy and control performance.
Cite
Text
Berto et al. "Neural Solvers for Fast and Accurate Numerical Optimal Control." International Conference on Learning Representations, 2022.Markdown
[Berto et al. "Neural Solvers for Fast and Accurate Numerical Optimal Control." International Conference on Learning Representations, 2022.](https://mlanthology.org/iclr/2022/berto2022iclr-neural/)BibTeX
@inproceedings{berto2022iclr-neural,
title = {{Neural Solvers for Fast and Accurate Numerical Optimal Control}},
author = {Berto, Federico and Massaroli, Stefano and Poli, Michael and Park, Jinkyoo},
booktitle = {International Conference on Learning Representations},
year = {2022},
url = {https://mlanthology.org/iclr/2022/berto2022iclr-neural/}
}