Sampled-data Systems: Stability, Contractivity and Single-iteration Suboptimal MPC

Yiting Chen, Francesco Bullo, Emiliano Dall’Anese · 2025

Read Sampled-data Systems: Stability, Contractivity and Single-iteration Suboptimal MPC by Yiting Chen, Francesco Bullo, Emiliano Dall’Anese (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-18336.txt.

Introduction

Chen, Bullo, and Dall'Anese analyze sampled-data systems in a 2025 paper. Such systems join a continuous-time part with a discrete-time part through sampling and hold steps. The setup models optimization-based controllers such as model predictive control, where the discrete part runs a set number of algorithm steps.

The authors define a reduced model that captures behavior as the sampling period shrinks. Their main result shows that if this reduced model is contractive, the full joined system is stable for small enough sampling periods. A small-gain argument covers the case where both parts are contractive. The results matter for real-time control, since stability can hold even with a single optimization step.

Catalog record

CREATORYiting Chen, Francesco Bullo, Emiliano Dall’Anese
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE109 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-18336

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Sampled-data Systems: Stability, Contractivity and Single-iteration Suboptimal MPC Yiting Chen, Francesco Bullo, Emiliano Dall’Anese Abstract This paper analyzes the stability of interconnected continuous-time (CT) and discrete-time (DT) systems coupled through sampling and zero-order hold mechanisms. The DT system updates its output at regular intervals T > 0 T>0 by applying an n n -fold composition of a given map. This setup is motivated by online and sampled-data implementations of optimization-based controllers – particularly model predictive control (MPC) – where the DT system models n n iterations of an algorithm approximating the solution of an optimization problem. We introduce the concept of a reduced model , defined as the limiting behavior of the sampled-data system as T → 0 + T\to 0^{+} and n → + ∞ n\to+\infty . Our main theoretical contribution establishes that when the reduced model is contractive, there exists a threshold duration T ⁡ ( n ) T(n) for each iteration count n n such that the CT-DT interconnection achieves exponential stability for all sampling periods T < T ⁡ ( n ) T<T(n) . Finally, under the stronger condition that both the CT and DT systems are contractive, we show exponential stability of their interconnection using a small-gain argument. Our theoretical results provide new insights into suboptimal MPC stability, showing that convergence guarantees hold even when using a single iteration of the optimization algorithm—a practically significant finding for real-time control applications. I Introduction This paper studies the interconnection of a continuous-time (CT) dynamical system with a discrete-time (DT) system, coupled via sampling and zero-order hold mechanisms. As illustrated in the block diagram in Fig. 1 (Left) , the DT system acquires samples of the state x regularly at intervals of length T , and updates its output z based on a map \mathsf{G}^{n}(x,z) ; in particular, \mathsf{G}^{n}(x,z) is defined as the n -fold composition of a given map \mathsf{G}(x,z) . In this setup, the goal of this paper is threefold: (i) First, inspired by the foundational work on two-time-scale CT systems [ 1 , 2 ] , we aim to formalize the notion of a reduced model (RM) associated with ( 1 ) by viewing T and n as parameters that induce a “time-scale separation”; a preview of our RM model is in Fig. 1 (Right) , where z^{*}(x):=\lim_{n\to+\infty}\mathsf{G}^{n}(x,z) for any fixed x . (ii) We examine how the properties of the RM, n , and T influence the stability of the interconnected system. These conditions are then compared with those derived from small-gain arguments. Our analysis builds on tools from contraction theory [ 3 , 4 ] , which provide a natural framework for studying stability of the interconnected system in Fig. 1 (Left) using Lipschitz constants for the CT and DT systems. For many systems of interest, these constants can be computed analytically or estimated numerically. Before moving onto a brief literature review, we first provide motivations for the overall setup. Fig. 1: (Left) Interconnected system considered in this paper, with a CT sub-system, a DT sub-system, and sampling and zero-order hold blocks. Here, \mathsf{G}^{n}(x,z) is defined as \mathsf{G}^{1}(x,z)=\mathsf{G}(x,z),\ldots,\mathsf{G}^{n}(x,z)=\mathsf{G}(x,\mathsf{G}^{n-1}(x,z)) . (Right) Reduced model. Here, for any x , z^{*}(x) is defined by z^{*}(x)=\mathsf{G}(x,z^{*}(x)) . Motivations . Our analysis is motivated by optimization- based control frameworks – which include, for example, model predictive control (MPC) [ 5 , 6 , 7 ] , Control Barrier Function (CBF) based control [ 8 , 9 ] and extensions including both CBF and Control Lyapunov Functions (CLFs) [ 10 ] , pointwise minimum-norm control [ 11 ] , and online feedback optimization [ 12 , 13 ] . These frameworks share a key characteristic: the control input is implicitly defined as the solution to an optimization problem that encodes performance objectives and system constraints. In our setup, the closed-loop system resulting from the interconnection of a plant with such an optimization-based controller is abstracted by the RM shown in Fig. 1 (Right) , where the map z^{*}(x) represents the control law obtained as the solution to the underlying optimization problem. Accordingly, the interconnected system depicted in Fig. 1 (Left) captures practical aspects of real-world implementations, where the system state x is sampled at discrete time intervals, and the computation of the control input z^{*}(x) is limited by computational constraints. Specifically, instead of solving the optimization problem to full convergence, only a finite number n of iterations of an algorithm are executed. We are particularly interested in applying our theoretical results to MPC problems, where the optimal solution map is often not available in closed form. This motivates the need for real-time MPC implementations [ 6 , 14 , 15 ] , as well as approaches such as instant MPC and sub-optimal MPC [ 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 ] . See also [ 24 ] for a broader perspective on online control and the recent work on receding horizon games in [ 25 ] . Connections to the literature . The study of two-time-scale CT systems dates back to the foundational works in singular perturbation theory [ 1 , 2 ] . A unified framework for analyzing stability and robustness in such systems was developed in [ 26 ] , while contraction-based approaches were explored in [ 27 , 28 ] . Contractivity of reduced-type dynamics has been studied in the context of nonlinear differential–algebraic equation systems in, e.g., [ 29 ] . Stability of coupled or networked CT dynamical systems was investigated in, e.g., [ 30 , 31 ] . Our analysis is inspired by these works, which prompted us to treat T and n as parameters that induce a time-scale separation. When considering the limiting case n=+\infty , T>0 , our framework encompasses classical sampled-data control systems [ 32 ] . The study of such systems has a rich literature, with numerous foundational contributions; representative works include [ 33 , 34 , 35 , 36 , 37 , 38 ] . Notably, in the context of linear time-invariant (LTI) systems, [ 34 ] establishes stability conditions based on the sampling period T , formulated through Linear Matrix Inequalities (LMIs). In the context of instant MPC and time-distributed MPC, existing stability analysis focuses on CT-CT interconnections [ 17 , 18 ] or DT-DT settings [ 19 , 20 , 21 ] . Contributions . Our contribution is threefold. (a) First, we formalize the notion of RM, which is defined as the limiting system for T\!\rightarrow\!0^{+} and n\!\rightarrow\!+\infty , modeling the case where the map z^{*}(x) is applied at all times. Then, we show our main result: if the DT system is contractive and the RM is strongly infinitesimally contractive, then for any n\in\mathbb{Z}_{>0} there exists a T(n) such that the interconnection between the CT and DT systems is globally exponentially stable (GES) for any T\!<\!T(n) . We offer a detailed expression for T(n) , which depends solely on n and on the Lipschitz constants of the RM and of the CT and DT systems. We also show that T(n) is strictly increasing with respect to n and it is bounded. (b) Under the assumption that the CT system and DT system are contractive, we show that the

Read the complete text