Year: 2019
Type: preprint
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.
Source: arXiv (Cornell University)
Authors: Roulet, Vincent, Srinivasa, Siddhartha, Drusvyatskiy, Dmitriy, Harchaoui, Zaid
Institution: University of Washington
Language: English
Subfield: Control and Systems Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
Awards: TRIPODS: Algorithms for Data Science: Complexity, Scalability, and Robustness., TRIPODS+X:RES: Safe Imitation Learning for Robotics, CAREER: Structure, Complexity, and Conditioning in Nonsmooth Optimization, NRI: Collaborative Research: Learning Deep Sensorimotor Policies for Shared Autonomy, CPS: Synergy: Collaborative Research: Learning control sharing strategies for assistive cyber-physical systems
OpenAlex ID: W2946431541