Skip to content

Iterative Linearized Control: Stable Algorithms and Complexity Guarantees

Authors
Vincent Roulet, Siddhartha Srinivasa, Dmitriy Drusvyatskiy, Zaid Harchaoui
Year
2019
Preprint
Identifiers
arXiv 1908.07615

Abstract

We examine popular gradient-based algorithms for nonlinear control in the light of the modern complexity analysis of first-order optimization algorithms. The examination reveals that the complexity bounds can be clearly stated in terms of calls to a computational oracle related to dynamic programming and implementable by gradient back-propagation using machine learning software libraries such as PyTorch or TensorFlow. Finally, we propose a regularized Gauss-Newton algorithm enjoying worst-case complexity bounds and improved convergence behavior in practice. The software library based on PyTorch is publicly available.

Knowledge Base

Map Tree Search

External

Document arXiv Abstract arXiv PDF arXiv HTML Google Scholar OpenAlex

Similar Papers