Episodic Linear Quadratic Regulators with Low-rank Transitions

Abstract

Linear Quadratic Regulators (LQR) achieve enormous successful real-world applications. Very recently, people have been focusing on efficient learning algorithms for LQRs when their dynamics are unknown. Existing results effectively learn to control the unknown system using number of episodes depending polynomially on the system parameters, including the ambient dimension of the states. These traditional approaches, however, become inefficient in common scenarios, e.g., when the states are high-resolution images. In this paper, we propose an algorithm that utilizes the intrinsic system low-rank structure for efficient learning. For problems of rank-$m$, our algorithm achieves a $K$-episode regret bound of order $\tilde{O}(m^{3/2}{K}^{1/2})$. Consequently, the sample complexity of our algorithm only depends on the rank, $m$, rather than the ambient dimension, $d$, which can be orders-of-magnitude larger.

References in corpus (7)

• Continuous control with deep reinforcement learning
• Contextual Decision Processes with Low Bellman Rank are PAC-Learnable
• Model-based RL in Contextual Decision Processes: PAC bounds and Exponential Improvements over Model-free Approaches
• Non-Asymptotic Analysis of Robust Control from Coarse-Grained Identification
• Thompson Sampling for Linear-Quadratic Control Problems
• Stochastic Low-Rank Bandits
• Stochastic Linear Bandits with Hidden Low Rank Structure

Cited by in corpus (1)

• Sparsity in Partially Controllable Linear Systems