An Algorithm for Learning Switched Linear Dynamics from DataDownload PDF

Published: 31 Oct 2022, Last Modified: 19 Jan 2023NeurIPS 2022 AcceptReaders: Everyone
Keywords: Control Systems, System Identification, Machine Learning, Hybrid Systems, Dynamical Systems, Ellipsoidal Algorithm, Separation Oracles
TL;DR: Solving the problem of identifying multiple dynamics from data using an algorithm that combines ideas from integer linear programming and ellipsoidal method.
Abstract: We present an algorithm for learning switched linear dynamical systems in discrete time from noisy observations of the system's full state or output. Switched linear systems use multiple linear dynamical modes to fit the data within some desired tolerance. They arise quite naturally in applications to robotics and cyber-physical systems. Learning switched systems from data is a NP-hard problem that is nearly identical to the $k$-linear regression problem of fitting $k > 1$ linear models to the data. A direct mixed-integer linear programming (MILP) approach yields time complexity that is exponential in the number of data points. In this paper, we modify the problem formulation to yield an algorithm that is linear in the size of the data while remaining exponential in the number of state variables and the desired number of modes. To do so, we combine classic ideas from the ellipsoidal method for solving convex optimization problems, and well-known oracle separation results in non-smooth optimization. We demonstrate our approach on a set of microbenchmarks and a few interesting real-world problems. Our evaluation suggests that the benefits of this algorithm can be made practical even against highly optimized off-the-shelf MILP solvers.
Supplementary Material: pdf
14 Replies