On the sample complexity of the linear quadratic regulator
From MaRDI portal
Publication:2194770
DOI10.1007/s10208-019-09426-yzbMath1447.49052arXiv1710.01688OpenAlexW2966348706WikidataQ127408103 ScholiaQ127408103MaRDI QIDQ2194770
Stephen Tu, Benjamin Recht, Horia Mania, Sarah Dean, Nikolai Matni
Publication date: 7 September 2020
Published in: Foundations of Computational Mathematics (Search for Journal in Brave)
Full work available at URL: https://arxiv.org/abs/1710.01688
optimal controlsystem identificationrobust controlreinforcement learningstatistical learning theorysystem level synthesis
Adaptive or robust stabilization (93D21) Identification in stochastic control theory (93E12) Robust stability (93D09) Linear-quadratic optimal control problems (49N10) Stochastic learning and adaptive control (93E35) Random matrices (algebraic aspects) (15B52)
Related Items
Model-free design of stochastic LQR controller from a primal-dual optimization perspective, Turnpike in optimal control of PDEs, ResNets, and beyond, Derivative-Free Methods for Policy Optimization: Guarantees for Linear Quadratic Systems, Control-oriented regularization for linear system identification, Determining optimal input-output properties: a data-driven approach, High-dimensional dynamic systems identification with additional constraints, Reinforcement Learning for Linear-Convex Models with Jumps via Stability Analysis of Feedback Controls, Analysis of the optimization landscape of Linear Quadratic Gaussian (LQG) control, Active Operator Inference for Learning Low-Dimensional Dynamical-System Models from Noisy Data, Quadratic Matrix Inequalities with Applications to Data-Based Control, Correct-By-Construction Exploration and Exploitation for Unknown Linear Systems Using Bilinear Optimization, Robustness to Incorrect System Models in Stochastic Control, Optimal Scheduling of Entropy Regularizer for Continuous-Time Linear-Quadratic Reinforcement Learning, Empirical risk minimization and complexity of dynamical models, Efficient Learning of Distributed Linear-Quadratic Control Policies, Policy Gradient Methods for the Noisy Linear Quadratic Regulator over a Finite Horizon, Trade-offs in learning controllers from noisy data, Unnamed Item, Low-complexity learning of linear quadratic regulators from noisy data, Data-driven control via Petersen's lemma, Bayesian frequentist bounds for machine learning and system identification, Controlled interacting particle algorithms for simulation-based reinforcement learning, Entropy Regularization for Mean Field Games with Learning
Uses Software
Cites Work
- Unnamed Item
- Unnamed Item
- Unnamed Item
- Unnamed Item
- Bootstrap methods: another look at the jackknife
- Rates of convergence for empirical processes of stationary mixing sequences
- A convex approach to robust \({\mathcal H}_{2}\) performance analysis
- Linear Thompson sampling revisited
- The complex structured singular value
- Nonasymptotic bounds for autoregressive time series modeling.
- The jackknife and bootstrap
- A formula for computation of the real stability radius
- Weak convergence and empirical processes. With applications to statistics
- Generalization bounds for non-stationary mixing processes
- Predictive Control for Linear and Hybrid Systems
- Analysis of Robust H2 Performance Using Multiplier Theory
- Robustness in the presence of mixed parametric uncertainty and unmodeled dynamics
- Control oriented system identification: a worst-case/deterministic approach in H/sub infinity /
- Modern Wiener-Hopf design of optimal controllers--Part II: The multivariable case
- Computational complexity of μ calculation
- System analysis via integral quadratic constraints
- Nonparametric estimation of transfer functions: rates of convergence and adaptation
- A Tutorial on Thompson Sampling
- Gradient Descent Learns Linear Dynamical Systems
- High-Dimensional Statistics
- A System-Level Approach to Controller Synthesis
- Finite sample properties of system identification methods
- Stability Analysis of Discrete-Time Infinite-Horizon Optimal Control With Discounted Cost
- Nonparametric risk bounds for time-series forecasting
- Positive trigonometric polynomials and signal processing applications
- The bootstrap and Edgeworth expansion
