A framework for randomized time-splitting in linear-quadratic optimal control
From MaRDI portal
Publication:2671271
General theory of random and stochastic dynamical systems (37H05) Simulation of dynamical systems (37M05) Optimality conditions for problems involving randomness (49K45) Linear-quadratic optimal control problems (49N10) Probabilistic methods, stochastic differential equations (65C99) Stability and convergence of numerical methods for ordinary differential equations (65L20)
Abstract: Inspired by the successes of stochastic algorithms in the training of deep neural networks and the simulation of interacting particle systems, we propose and analyze a framework for randomized time-splitting in linear-quadratic optimal control. In our proposed framework, the linear dynamics of the original problem is replaced by a randomized dynamics. To obtain the randomized dynamics, the system matrix is split into simpler submatrices and the time interval of interest is split into subintervals. The randomized dynamics is then found by selecting randomly one or more submatrices in each subinterval. We show that the dynamics, the minimal values of the cost functional, and the optimal control obtained with the proposed randomized time-splitting method converge in expectation to their analogues in the original problem when the time grid is refined. The derived convergence rates are validated in several numerical experiments. Our numerical results also indicate that the proposed method can lead to a reduction in computational cost for the simulation and optimal control of large-scale linear dynamical systems.
Recommendations
- Numerical solution of the finite horizon stochastic linear quadratic control problem
- A distributed stochastic approximation algorithm for stochastic LQ control with unknown uncertainty
- Controlled interacting particle algorithms for simulation-based reinforcement learning
- Error analysis of a discretization for stochastic linear quadratic control problems governed by SDEs
- A numerical method for solving stochastic optimal control problems with linear control
Cites work
- A proposal on machine learning via dynamical systems
- A Random-Batch Monte Carlo Method for Many-Body Systems with Singular Kernels
- A splitting method for the nonlinear Schrödinger equation
- An introduction to domain decomposition methods. Algorithms, theory, and parallel implementation
- An introduction to probability and statistics
- Control under constraints for multi-dimensional reaction-diffusion monostable and bistable equations
- Controllability of a one-dimensional fractional heat equation: theoretical and numerical aspects
- Convergence of the random batch method for interacting particles with disparate species and weights
- Crank--Nicolson schemes for optimal control problems with evolution equations
- Deep learning as optimal control problems: models and numerical methods
- Dimension splitting for evolution equations
- Generalization of the Trotter-Lie formula
- scientific article; zbMATH DE number 3972625 (Why is no real title available?)
- scientific article; zbMATH DE number 3721567 (Why is no real title available?)
- scientific article; zbMATH DE number 3256013 (Why is no real title available?)
- Model predictive control with random batch methods for a guiding problem
- Nonlinear model predictive control. Theory and algorithms
- On error estimates for the Trotter-Kato product formula
- On the Product of Semi-Groups of Operators
- On the stabilizability of the Burgers equation by receding horizon control
- Optimization methods for large-scale machine learning
- Perturbations of nonlinear systems of differential equations
- Random batch methods (RBM) for interacting particle systems
- Receding horizon control for the stabilization of the wave equation
- Stability of exponential operator splitting methods for noncontractive semigroups
- Turnpike in Lipschitz-nonlinear optimal control
- Unconstrained model predictive control and suboptimality estimates for nonlinear continuous-time systems
Cited in
(5)- Resource-aware time-optimal control with multiple sparsity measures
- Stabilizability of parabolic equations by switching controls based on point actuators
- Random sampling-based gradient descent method for optimal control problems with variance reduction
- Mini-batch descent in semiflows
- Output-based receding horizon stabilizing control for linear parabolic equations
This page was built for publication: A framework for randomized time-splitting in linear-quadratic optimal control
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2671271)
