Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

On the Convergence of Overlapping Schwarz Decomposition for Nonlinear Optimal Control

Journal Article · · IEEE Transactions on Automatic Control
 [1];  [2];  [3];  [4]
  1. Univ. of Chicago, IL (United States)
  2. Univ. of Wisconsin, Madison, WI (United States)
  3. Argonne National Laboratory (ANL), Argonne, IL (United States); Univ. of Chicago, IL (United States)
  4. Univ. of Wisconsin, Madison, WI (United States); Argonne National Laboratory (ANL), Argonne, IL (United States)
Here, we study the convergence properties of an overlapping Schwarz decomposition algorithm for solving nonlinear optimal control problems (OCPs). The algorithm decomposes the time domain into a set of overlapping subdomains, and solves all subproblems defined over subdomains in parallel. The convergence is attained by updating primal-dual information at the boundaries of overlapping subdomains. We show that the algorithm exhibits local linear convergence, and that the convergence rate improves exponentially with the overlap size. We also establish global convergence results for a general quadratic programming, which enables the application of the Schwarz scheme inside second-order optimization algorithms (e.g., sequential quadratic programming). The theoretical foundation of our convergence analysis is a sensitivity result of nonlinear OCPs, which we call "exponential decay of sensitivity" (EDS). Intuitively, EDS states that the impact of perturbations at domain boundaries (i.e., initial and terminal time) on the solution decays exponentially as one moves into the domain. Here, we expand a previous analysis available in the literature by showing that EDS holds for both primal and dual solutions of nonlinear OCPs, under uniform second-order sufficient condition, controllability condition, and boundedness condition. We conduct experiments with a quadrotor motion planning problem and a partial differential equations (PDE) control problem to validate our theory, and show that the approach is significantly more efficient than alternating direction method of multipliers and as efficient as the centralized interior-point solver.
Research Organization:
Argonne National Laboratory (ANL), Argonne, IL (United States)
Sponsoring Organization:
USDOE Office of Science (SC), Advanced Scientific Computing Research (ASCR); National Science Foundation (NSF)
Grant/Contract Number:
AC02-06CH11357
OSTI ID:
2333647
Journal Information:
IEEE Transactions on Automatic Control, Journal Name: IEEE Transactions on Automatic Control Journal Issue: 11 Vol. 67; ISSN 0018-9286
Publisher:
IEEECopyright Statement
Country of Publication:
United States
Language:
English

References (45)

Optimal infinite-horizon feedback laws for a general class of constrained discrete-time systems: Stability and moving-horizon approximations journal May 1988
A parallel Newton-type method for nonlinear model predictive control journal November 2019
A Parallel Decomposition Scheme for Solving Long-Horizon Optimal Control Problems conference December 2019
A flying inverted pendulum conference May 2011
Exponentially Accurate Temporal Decomposition for Long-Horizon Linear-Quadratic Dynamic Optimization journal January 2018
Exponential Decay in the Sensitivity Analysis of Nonlinear Dynamic Programming journal January 2020
Perturbed Kuhn-Tucker points and rates of convergence for a class of nonlinear-programming algorithms journal December 1974
On the implementation of an interior-point filter line-search algorithm for large-scale nonlinear programming journal April 2005
Global Convergence of ADMM in Nonconvex Nonsmooth Optimization journal June 2018
Solution of discrete-time optimal control problems on parallel computers journal December 1990
A survey of industrial model predictive control technology journal July 2003
A continuation/GMRES method for fast computation of nonlinear receding horizon control journal April 2004
The advanced-step NMPC controller: Optimality, stability and robustness journal January 2009
Parallel cyclic reduction decomposition for dynamic optimization problems journal January 2019
Benchmarking ADMM in nonconvex NLPs journal November 2018
New Architectures for Hierarchical Predictive Control journal January 2016
A Stochastic Dual Dynamic Programming Framework for Multiscale MPC journal January 2018
Graph-Based Modeling and Decomposition of Energy Infrastructures journal January 2021
Stochastic model predictive control for central HVAC plants journal June 2020
Temporal Decomposition Scheme for Nonlinear Multisite Production Planning and Distribution Models journal May 2003
Application of Interior-Point Methods to Model Predictive Control journal December 1998
Nominal stability of real-time iteration scheme for nonlinear model predictive control journal May 2005
Abstract nonlinear sensitivity and turnpike analysis and an application to semilinear parabolic PDEs journal January 2021
Benchmarking large-scale distributed convex quadratic programming algorithms journal May 2014
Non quadratic smooth model of fatigue for optimal fatigue-oriented individual pitch control journal September 2020
Diffusing-Horizon Model Predictive Control journal January 2023
Efficient implementation of the Riccati recursion for solving linear-quadratic control problems conference August 2013
Temporal Lagrangian decomposition of model predictive control for hybrid systems conference January 2004
A parallel structure exploiting factorization algorithm with applications to Model Predictive Control conference December 2015
Nonlinear programming strategies on high-performance computers conference December 2015
Parallelizing LQR Computation Through Endpoint-Explicit Riccati Recursion conference December 2019
Distributed Receding Horizon Control of Dynamically Coupled Nonlinear Systems journal July 2007
An Augmented Lagrangian Filter Method for Real-Time Embedded Optimization journal December 2017
Superconvergence of Online Optimization for Model Predictive Control journal March 2023
Predictive Active Steering Control for Autonomous Vehicle Systems journal May 2007
Neural-network predictive control for nonlinear dynamic systems with time-delay journal March 2003
General nonlinear modal representation of large scale power systems journal August 2003
Real-Time Nonlinear Optimization as a Generalized Equation journal January 2010
Convergence Analysis of Alternating Direction Method of Multipliers for a Family of Nonconvex Problems journal January 2016
A Sparsity Preserving Convexification Procedure for Indefinite Quadratic Programs Arising in Direct Optimal Control journal January 2017
Efficient Model Predictive Control for Parabolic PDEs with Goal Oriented Error Estimation journal February 2022
A Real-Time Iteration Scheme for Nonlinear Optimization in Optimal Feedback Control journal January 2005
Distributed Optimization and Statistical Learning via the Alternating Direction Method of Multipliers book January 2011
Handling Long Horizons in MPC: A Stochastic Programming Approach conference June 2018
An Ô (log N ) Parallel Algorithm for Newton Step Computation in Model Predictive Control journal January 2014

Cited By (3)

Controllability and Observability Imply Exponential Decay of Sensitivity in Dynamic Optimization journal January 2021
Abstract nonlinear sensitivity and turnpike analysis and an application to semilinear parabolic PDEs journal January 2021
Exponential Decay of Sensitivity in Graph-Structured Nonlinear Programs journal May 2022

Figures / Tables (7)


Similar Records

Schwarz method for earthquake source dynamics
Journal Article · Tue Apr 01 00:00:00 EDT 2008 · Journal of Computational Physics · OSTI ID:21028319

Overlapping Schwarz for Nonlinear Problems. An Element Agglomeration Nonlinear Additive Schwarz Preconditioned Newton Method for Unstructured Finite Element Problems
Journal Article · Wed Feb 09 23:00:00 EST 2005 · Application of Mathematics, vol. 50, no. 3, July 12, 2005, pp. 247-275 · OSTI ID:936450