Pub Date : 2025-11-13DOI: 10.1109/LCSYS.2025.3632307
Yanan Zhu;Wenwu Yu;Guanghui Wen
This letter provides the first rigorous theoretical analysis for the Distributed Fenchel Dual Gradient (DFDG) algorithm, a continuous-time method for solving distributed convex optimization problems with local set constraints over digraphs. The DFDG algorithm, originally proposed in our prior work (Zhu et al., 2020), transforms the primal problem into its Fenchel dual and solves it using a two-time-scale dynamical system. This letter provides a more comprehensive explanation of the algorithm’s design mechanism and formally establishes its convergence properties. Under strong convexity and Lipschitz continuity assumptions, Lyapunov stability theory is employed to prove the asymptotic convergence to the optimal solutions of both the primal and its dual problems. This analysis provides rigorous guarantees for a class of dual-based algorithms over digraphs, filling a critical gap in the existing literature.
这封信为分布式Fenchel对偶梯度（DFDG）算法提供了第一个严格的理论分析，这是一种连续时间方法，用于解决有向图上具有局部集约束的分布式凸优化问题。DFDG算法最初是在我们之前的工作中提出的（Zhu et al., 2020），它将原始问题转换为Fenchel对偶，并使用双时间尺度动力系统来解决它。这封信提供了一个更全面的解释算法的设计机制，并正式建立了其收敛性质。在强凸性和Lipschitz连续性假设下，利用Lyapunov稳定性理论证明了该问题的最优解及其对偶问题的渐近收敛性。这种分析为有向图上的一类基于双重的算法提供了严格的保证，填补了现有文献中的一个关键空白。
{"title":"Distributed Fenchel Dual Gradient Algorithm for Constrained Convex Optimization Over Digraphs","authors":"Yanan Zhu;Wenwu Yu;Guanghui Wen","doi":"10.1109/LCSYS.2025.3632307","DOIUrl":"https://doi.org/10.1109/LCSYS.2025.3632307","url":null,"abstract":"This letter provides the first rigorous theoretical analysis for the Distributed Fenchel Dual Gradient (DFDG) algorithm, a continuous-time method for solving distributed convex optimization problems with local set constraints over digraphs. The DFDG algorithm, originally proposed in our prior work (Zhu et al., 2020), transforms the primal problem into its Fenchel dual and solves it using a two-time-scale dynamical system. This letter provides a more comprehensive explanation of the algorithm’s design mechanism and formally establishes its convergence properties. Under strong convexity and Lipschitz continuity assumptions, Lyapunov stability theory is employed to prove the asymptotic convergence to the optimal solutions of both the primal and its dual problems. This analysis provides rigorous guarantees for a class of dual-based algorithms over digraphs, filling a critical gap in the existing literature.","PeriodicalId":37235,"journal":{"name":"IEEE Control Systems Letters","volume":"9 ","pages":"2537-2542"},"PeriodicalIF":2.0,"publicationDate":"2025-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145560763","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-07DOI: 10.1109/LCSYS.2025.3630241
Hesham Abdelfattah;Sameh A. Eisa;Peter Stechlinski
In this letter, we extend the sensitivity-based rank condition (SERC) test for local observability to another class of systems, namely smooth and nonsmooth differential-algebraic equation (DAE) systems of index-1. The newly introduced test for DAEs, which we call the lexicographic SERC (L-SERC) observability test, utilizes the theory of lexicographic differentiation to compute sensitivity information. Moreover, the newly introduced L-SERC observability test can judges which states are observable and which are not. Additionally, we introduce a novel sensitivity-based extended Kalman filter (S-EKF) algorithm for state estimation, applicable to both smooth and nonsmooth DAE systems. Finally, we apply the newly developed S-EKF to estimate the states of a wind turbine power system model.
{"title":"Observability and State Estimation for Smooth and Nonsmooth Differential Algebraic Equation Systems","authors":"Hesham Abdelfattah;Sameh A. Eisa;Peter Stechlinski","doi":"10.1109/LCSYS.2025.3630241","DOIUrl":"https://doi.org/10.1109/LCSYS.2025.3630241","url":null,"abstract":"In this letter, we extend the sensitivity-based rank condition (SERC) test for local observability to another class of systems, namely smooth and nonsmooth differential-algebraic equation (DAE) systems of index-1. The newly introduced test for DAEs, which we call the lexicographic SERC (L-SERC) observability test, utilizes the theory of lexicographic differentiation to compute sensitivity information. Moreover, the newly introduced L-SERC observability test can judges which states are observable and which are not. Additionally, we introduce a novel sensitivity-based extended Kalman filter (S-EKF) algorithm for state estimation, applicable to both smooth and nonsmooth DAE systems. Finally, we apply the newly developed S-EKF to estimate the states of a wind turbine power system model.","PeriodicalId":37235,"journal":{"name":"IEEE Control Systems Letters","volume":"9 ","pages":"2507-2512"},"PeriodicalIF":2.0,"publicationDate":"2025-11-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145510102","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Hierarchical reinforcement learning (RL) aims to improve sample efficiency by decomposing complex long-horizon tasks into fast low-level myopic and slower high-level non-myopic subtasks. However, the unilateral nested policy structure in current goal-conditioned hierarchical RL (HRL) methods sets subgoals at the high level without considering feedback from the low level, which significantly degrades the performance of high-level subgoal generation and sampling efficiency. Hindsight action relabeling further weakens subgoal settings by submitting to low-level reachability. Inspired by feedback control of dynamic systems, we present Feedback for Improved HRL with Timed Subgoals (FIHTS), a mechanism allowing feedback control of subgoal generation for improved HRL. Unlike current HRL, FIHTS enables both the high level to set subgoals and the low level to receive rewards based on subgoal achievement. Our experiments in various challenging dynamic RL environments show that our FIHTS method achieves higher success rates with higher sample efficiency than existing subgoal-based HRL methods.
{"title":"Feedback for Improved Hierarchical Reinforcement Learning With Timed Subgoals","authors":"Yajie Bao;Dan Shen;Genshe Chen;Hao Xu;Samson Badlia;Simon Khan;Erik Blasch;Khanh Pham","doi":"10.1109/LCSYS.2025.3629008","DOIUrl":"https://doi.org/10.1109/LCSYS.2025.3629008","url":null,"abstract":"Hierarchical reinforcement learning (RL) aims to improve sample efficiency by decomposing complex long-horizon tasks into fast low-level myopic and slower high-level non-myopic subtasks. However, the unilateral nested policy structure in current goal-conditioned hierarchical RL (HRL) methods sets subgoals at the high level without considering feedback from the low level, which significantly degrades the performance of high-level subgoal generation and sampling efficiency. Hindsight action relabeling further weakens subgoal settings by submitting to low-level reachability. Inspired by feedback control of dynamic systems, we present Feedback for Improved HRL with Timed Subgoals (FIHTS), a mechanism allowing feedback control of subgoal generation for improved HRL. Unlike current HRL, FIHTS enables both the high level to set subgoals and the low level to receive rewards based on subgoal achievement. Our experiments in various challenging dynamic RL environments show that our FIHTS method achieves higher success rates with higher sample efficiency than existing subgoal-based HRL methods.","PeriodicalId":37235,"journal":{"name":"IEEE Control Systems Letters","volume":"9 ","pages":"2501-2506"},"PeriodicalIF":2.0,"publicationDate":"2025-11-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145510104","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-10-30DOI: 10.1109/LCSYS.2025.3627045
Soham Mukherjee;Mrityunjoy Chakraborty
A decentralized stochastic optimization problem is considered, in which a network of nodes collaborate over noisy communication links to minimize a global objective function. Recently, the authors in Mukherjee and Chakraborty (2025) incorporated a skipping technique in the Noisy Consensus + Stochastic Gradient Descent (SGD) framework to address this problem, and showed that their proposed skipping technique helps improve the ${mathrm { O}}text {(}1/epsilon ^{3}text {)}$ computational complexity obtained in previous works to ${mathrm { O}}text {(}1/epsilon ^{2}text {)}$ under a general smoothness assumption. In this letter, we consider the algorithm proposed in Mukherjee and Chakraborty (2025) to show how the computational complexity can be further improved to ${mathrm { O}}text {(}{mathrm { log}}text {(}1/epsilon text {)}/epsilon text {)}$ when the Polyak-Lojasiewicz (PL) condition is satisfied in addition to the smoothness assumption. The obtained ${mathrm { O}}text {(}{mathrm { log}}text {(}1/epsilon text {)}/epsilon text {)}$ rate in the current work is also an Improvement over the ${mathrm { O}}text {(}1/epsilon ^{2}text {)}$ rate obtained in previous works under the strong-convexity assumption (which is known to be stricter than the PL condition), and matches the $Omega text {(}1/epsilon text {)}$ lower bound for the number of stochastic gradient computations for the considered problem class up to an extra ${mathrm { log}}text {(}1/epsilon text {)}$ factor. Last but not least, the ${mathrm { O}}text {(}{mathrm { log}}text {(}1/epsilon text {)}/epsilon text {)}$ computational cost is achieved while retaining the ${mathrm { O}}text {(}{mathrm { log}}^{2}text {(}1/epsilon text {)}/epsilon ^{2}text {)}$ rate for the number of iterations and communication rounds, which is at par with the results obtained in previous works which consider strong-convexity, up to logarithmic factors. A numerical experiment is conducted corroborate theoretical results.
研究了一个分散的随机优化问题，其中节点网络在有噪声的通信链路上协作以最小化全局目标函数。最近，Mukherjee和Chakraborty（2025）的作者在噪声一致+随机梯度下降（SGD）框架中引入了一种跳过技术来解决这个问题，并表明他们提出的跳过技术有助于在一般平滑假设下将先前工作中获得的${mathrm { O}}text {(}1/epsilon ^{3}text {)}$计算复杂度提高到${mathrm { O}}text {(}1/epsilon ^{2}text {)}$。在这封信中，我们考虑了Mukherjee和Chakraborty（2025）中提出的算法，以展示在平滑假设之外满足Polyak-Lojasiewicz （PL）条件时如何将计算复杂度进一步提高到${mathrm { O}}text {(}{mathrm { log}}text {(}1/epsilon text {)}/epsilon text {)}$。在当前工作中获得的${mathrm { O}}text {(}{mathrm { log}}text {(}1/epsilon text {)}/epsilon text {)}$比率也是在强凸性假设下（已知比PL条件更严格）的先前工作中获得的${mathrm { O}}text {(}1/epsilon ^{2}text {)}$比率的改进，并且与所考虑的问题类别的随机梯度计算次数的$Omega text {(}1/epsilon text {)}$下界匹配到一个额外的${mathrm { log}}text {(}1/epsilon text {)}$因子。最后但并非最不重要的是，${mathrm { O}}text {(}{mathrm { log}}text {(}1/epsilon text {)}/epsilon text {)}$计算成本是在保持迭代次数和通信轮数的${mathrm { O}}text {(}{mathrm { log}}^{2}text {(}1/epsilon text {)}/epsilon ^{2}text {)}$速率的同时实现的，这与之前考虑强凸性的工作中获得的结果相当，直到对数因子。数值实验验证了理论结果。
{"title":"Toward O(log(1/ϵ)/ϵ) Computational Complexity for PL Functions in Decentralized Stochastic Optimization With Communication Noise","authors":"Soham Mukherjee;Mrityunjoy Chakraborty","doi":"10.1109/LCSYS.2025.3627045","DOIUrl":"https://doi.org/10.1109/LCSYS.2025.3627045","url":null,"abstract":"A decentralized stochastic optimization problem is considered, in which a network of nodes collaborate over noisy communication links to minimize a global objective function. Recently, the authors in Mukherjee and Chakraborty (2025) incorporated a skipping technique in the Noisy Consensus + Stochastic Gradient Descent (SGD) framework to address this problem, and showed that their proposed skipping technique helps improve the <inline-formula> <tex-math>${mathrm { O}}text {(}1/epsilon ^{3}text {)}$ </tex-math></inline-formula> computational complexity obtained in previous works to <inline-formula> <tex-math>${mathrm { O}}text {(}1/epsilon ^{2}text {)}$ </tex-math></inline-formula> under a general smoothness assumption. In this letter, we consider the algorithm proposed in Mukherjee and Chakraborty (2025) to show how the computational complexity can be further improved to <inline-formula> <tex-math>${mathrm { O}}text {(}{mathrm { log}}text {(}1/epsilon text {)}/epsilon text {)}$ </tex-math></inline-formula> when the Polyak-Lojasiewicz (PL) condition is satisfied in addition to the smoothness assumption. The obtained <inline-formula> <tex-math>${mathrm { O}}text {(}{mathrm { log}}text {(}1/epsilon text {)}/epsilon text {)}$ </tex-math></inline-formula> rate in the current work is also an Improvement over the <inline-formula> <tex-math>${mathrm { O}}text {(}1/epsilon ^{2}text {)}$ </tex-math></inline-formula> rate obtained in previous works under the strong-convexity assumption (which is known to be stricter than the PL condition), and matches the <inline-formula> <tex-math>$Omega text {(}1/epsilon text {)}$ </tex-math></inline-formula> lower bound for the number of stochastic gradient computations for the considered problem class up to an extra <inline-formula> <tex-math>${mathrm { log}}text {(}1/epsilon text {)}$ </tex-math></inline-formula> factor. Last but not least, the <inline-formula> <tex-math>${mathrm { O}}text {(}{mathrm { log}}text {(}1/epsilon text {)}/epsilon text {)}$ </tex-math></inline-formula> computational cost is achieved while retaining the <inline-formula> <tex-math>${mathrm { O}}text {(}{mathrm { log}}^{2}text {(}1/epsilon text {)}/epsilon ^{2}text {)}$ </tex-math></inline-formula> rate for the number of iterations and communication rounds, which is at par with the results obtained in previous works which consider strong-convexity, up to logarithmic factors. A numerical experiment is conducted corroborate theoretical results.","PeriodicalId":37235,"journal":{"name":"IEEE Control Systems Letters","volume":"9 ","pages":"2477-2482"},"PeriodicalIF":2.0,"publicationDate":"2025-10-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145455747","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-10-30DOI: 10.1109/LCSYS.2025.3626794
Jun Chen;Feng Lin
This letter proposes a new framework to capture detectability property in stochastic discrete-event systems. A new notion, name Partition-based Detectability or P-Detectability, is proposed based on partitions of the system state space, rather than the state space itself. In other words, the proposed P-Detectability focuses on the system capability to detect certain state group from other state groups, while ignoring the ambiguity between individual states within the same state group. As a consequence, the proposed P-Detectability allows users to define customized public and cover to ignore irrelevant ambiguity. Compared to existing notions such as A-Detectability and A-Diagnosabiltiy, the proposed notion is shown to be more general. A necessary and sufficient condition to verify P-Detectability, together with a testing algorithm, are developed.
{"title":"A General Framework for Detectability in Stochastic Discrete-Event Systems","authors":"Jun Chen;Feng Lin","doi":"10.1109/LCSYS.2025.3626794","DOIUrl":"https://doi.org/10.1109/LCSYS.2025.3626794","url":null,"abstract":"This letter proposes a new framework to capture detectability property in stochastic discrete-event systems. A new notion, name Partition-based Detectability or P-Detectability, is proposed based on partitions of the system state space, rather than the state space itself. In other words, the proposed P-Detectability focuses on the system capability to detect certain state group from other state groups, while ignoring the ambiguity between individual states within the same state group. As a consequence, the proposed P-Detectability allows users to define customized public and cover to ignore irrelevant ambiguity. Compared to existing notions such as A-Detectability and A-Diagnosabiltiy, the proposed notion is shown to be more general. A necessary and sufficient condition to verify P-Detectability, together with a testing algorithm, are developed.","PeriodicalId":37235,"journal":{"name":"IEEE Control Systems Letters","volume":"9 ","pages":"2465-2470"},"PeriodicalIF":2.0,"publicationDate":"2025-10-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145455762","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This letter presents a novel finite- and fixed-time convergent framework for solving constrained convex optimization problems using safe gradient flow dynamics. In the existing literature, finite- and fixed-time gradient flow methods have primarily addressed equality constraints by formulating Lagrangian-based gradient dynamics or a scaled augmented Lagrangian formulation. In contrast, the proposed method augments the gradient flow with constraint-driven control inputs derived using control barrier function (CBF) techniques, enabling simultaneous handling of equality and inequality constraints without relying on penalty function formulations. This formulation treats decision variables as system states and Lagrange multipliers as control inputs, enabling the feasible set to act as an invariant safe set throughout the system’s evolution. The approach ensures strict feasibility and guarantees convergence to the optimal solution in both finite- and fixed- time. Rigorous Lyapunov-based analysis establishes theoretical stability guarantees, and numerical simulations demonstrate rapid, exact convergence for problems involving both equality and inequality constraints.
{"title":"Finite- and Fixed-Time Gradient Flows for Constrained Optimization via Control Barrier Functions","authors":"Baby Diana;Shyam Kamal;Sandip Ghosh;Thach Ngoc Dinh","doi":"10.1109/LCSYS.2025.3627044","DOIUrl":"https://doi.org/10.1109/LCSYS.2025.3627044","url":null,"abstract":"This letter presents a novel finite- and fixed-time convergent framework for solving constrained convex optimization problems using safe gradient flow dynamics. In the existing literature, finite- and fixed-time gradient flow methods have primarily addressed equality constraints by formulating Lagrangian-based gradient dynamics or a scaled augmented Lagrangian formulation. In contrast, the proposed method augments the gradient flow with constraint-driven control inputs derived using control barrier function (CBF) techniques, enabling simultaneous handling of equality and inequality constraints without relying on penalty function formulations. This formulation treats decision variables as system states and Lagrange multipliers as control inputs, enabling the feasible set to act as an invariant safe set throughout the system’s evolution. The approach ensures strict feasibility and guarantees convergence to the optimal solution in both finite- and fixed- time. Rigorous Lyapunov-based analysis establishes theoretical stability guarantees, and numerical simulations demonstrate rapid, exact convergence for problems involving both equality and inequality constraints.","PeriodicalId":37235,"journal":{"name":"IEEE Control Systems Letters","volume":"9 ","pages":"2483-2488"},"PeriodicalIF":2.0,"publicationDate":"2025-10-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145510103","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-10-28DOI: 10.1109/LCSYS.2025.3626276
Shenglu Wang;Kairui Feng;Mengqi Xue;Yue Song
The chance-constrained optimal power flow (CC-OPF) essentially finds the low-cost generation dispatch scheme ensuring operational constraints are met with a specified probability, termed the security level. While the security level is a crucial input parameter, how it shapes the CC-OPF feasibility boundary has not been revealed. Changing the security level from a parameter to a decision variable, this letter proposes a security maximization approach based on the chance constrained DC-OPF model, termed inverse CC-OPF (ICC-OPF), that seeks the maximum security level achievable by the system. To efficiently solve the ICC-OPF, we design a Newton-Raphson-like iteration algorithm leveraging the duality-based sensitivity analysis of an associated surrogate problem. Numerical experiments validate the proposed approach, revealing complex feasibility boundaries for security levels that underscore the importance of coordinating security levels across multiple chance constraints.
{"title":"Power Flow Security Maximization via Inverse Chance Constrained Optimization","authors":"Shenglu Wang;Kairui Feng;Mengqi Xue;Yue Song","doi":"10.1109/LCSYS.2025.3626276","DOIUrl":"https://doi.org/10.1109/LCSYS.2025.3626276","url":null,"abstract":"The chance-constrained optimal power flow (CC-OPF) essentially finds the low-cost generation dispatch scheme ensuring operational constraints are met with a specified probability, termed the security level. While the security level is a crucial input parameter, how it shapes the CC-OPF feasibility boundary has not been revealed. Changing the security level from a parameter to a decision variable, this letter proposes a security maximization approach based on the chance constrained DC-OPF model, termed inverse CC-OPF (ICC-OPF), that seeks the maximum security level achievable by the system. To efficiently solve the ICC-OPF, we design a Newton-Raphson-like iteration algorithm leveraging the duality-based sensitivity analysis of an associated surrogate problem. Numerical experiments validate the proposed approach, revealing complex feasibility boundaries for security levels that underscore the importance of coordinating security levels across multiple chance constraints.","PeriodicalId":37235,"journal":{"name":"IEEE Control Systems Letters","volume":"9 ","pages":"2489-2494"},"PeriodicalIF":2.0,"publicationDate":"2025-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145455765","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-10-28DOI: 10.1109/LCSYS.2025.3626272
Mohammad Jeddi;Mohammad Hossein Khademi;Ali Khaki-Sedigh
Data-driven control methodologies design controllers directly from data without explicit system models. Among these methods, the Virtual Reference Feedback Tuning (VRFT) method uses offline data collected from an unknown system. However, challenges arise in selecting an appropriate reference model and in handling non-minimum phase (NMP) transmission zeros. This letter extends an optimal reference-model selection algorithm from single-input single-output (SISO) to multivariable systems and introduces a one-shot method for identifying and incorporating NMP transmission zeros. An appropriate cost function is formulated for multivariable systems, and evolutionary optimization techniques are applied for reference-model selection. Simulation results validate the effectiveness of the proposed approach, contributing to improved data-driven control design for multivariable systems.
{"title":"Data-Driven Control Based on Virtual Reference Feedback Tuning With Optimal Reference Model for Non-Minimum Phase Multivariable Systems","authors":"Mohammad Jeddi;Mohammad Hossein Khademi;Ali Khaki-Sedigh","doi":"10.1109/LCSYS.2025.3626272","DOIUrl":"https://doi.org/10.1109/LCSYS.2025.3626272","url":null,"abstract":"Data-driven control methodologies design controllers directly from data without explicit system models. Among these methods, the Virtual Reference Feedback Tuning (VRFT) method uses offline data collected from an unknown system. However, challenges arise in selecting an appropriate reference model and in handling non-minimum phase (NMP) transmission zeros. This letter extends an optimal reference-model selection algorithm from single-input single-output (SISO) to multivariable systems and introduces a one-shot method for identifying and incorporating NMP transmission zeros. An appropriate cost function is formulated for multivariable systems, and evolutionary optimization techniques are applied for reference-model selection. Simulation results validate the effectiveness of the proposed approach, contributing to improved data-driven control design for multivariable systems.","PeriodicalId":37235,"journal":{"name":"IEEE Control Systems Letters","volume":"9 ","pages":"2459-2464"},"PeriodicalIF":2.0,"publicationDate":"2025-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145455745","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-10-27DOI: 10.1109/LCSYS.2025.3625957
Charis Stamouli;Leonardo F. Toso;Anastasios Tsiamis;George J. Pappas;James Anderson
We analyze the performance of policy gradient in multitask linear quadratic regulation (LQR), where the system and cost parameters differ across tasks. The main goal of multitask LQR is to find a controller with satisfactory performance on every task. Prior analyses on relevant contexts fail to capture closed-loop task similarities, resulting in conservative performance guarantees. To account for such similarities, we propose bisimulation-based measures of task heterogeneity. Our measures employ new bisimulation functions to bound the cost gradient distance between a pair of tasks in closed loop with a common stabilizing controller. Employing these measures, we derive suboptimality bounds for both the multitask optimal controller and the asymptotic policy gradient controller with respect to each of the tasks. We further provide conditions under which the policy gradient iterates remain stabilizing for every system. For multiple random sets of certain tasks, we observe that our bisimulation-based measures improve upon baseline measures of task heterogeneity dramatically.
{"title":"Policy Gradient Bounds in Multitask LQR","authors":"Charis Stamouli;Leonardo F. Toso;Anastasios Tsiamis;George J. Pappas;James Anderson","doi":"10.1109/LCSYS.2025.3625957","DOIUrl":"https://doi.org/10.1109/LCSYS.2025.3625957","url":null,"abstract":"We analyze the performance of policy gradient in multitask linear quadratic regulation (LQR), where the system and cost parameters differ across tasks. The main goal of multitask LQR is to find a controller with satisfactory performance on every task. Prior analyses on relevant contexts fail to capture closed-loop task similarities, resulting in conservative performance guarantees. To account for such similarities, we propose bisimulation-based measures of task heterogeneity. Our measures employ new bisimulation functions to bound the cost gradient distance between a pair of tasks in closed loop with a common stabilizing controller. Employing these measures, we derive suboptimality bounds for both the multitask optimal controller and the asymptotic policy gradient controller with respect to each of the tasks. We further provide conditions under which the policy gradient iterates remain stabilizing for every system. For multiple random sets of certain tasks, we observe that our bisimulation-based measures improve upon baseline measures of task heterogeneity dramatically.","PeriodicalId":37235,"journal":{"name":"IEEE Control Systems Letters","volume":"9 ","pages":"2495-2500"},"PeriodicalIF":2.0,"publicationDate":"2025-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145455838","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-10-27DOI: 10.1109/LCSYS.2025.3626269
Mengxiang Zeng;Peng Lin
Most of the existing studies on position constraints integrate the constraint conditions into the agent dynamics. In contrast, the active constraints considered in this letter are implemented through control algorithms, ensuring that the agents’ states remain within their constraint sets without altering the system dynamics. Therefore, this letter explores the active-constrained consensus problem with nonconvex velocity and convex position constraints. The position constraints are fundamentally different from the velocity constraints, because the velocity constraint sets contain the origin, whereas the position constraint sets may not. The key difficulty is how to deal with the coupling between these two fundamentally different constraints. By performing a series of model transformations and utilizing the convexity of the system, it is demonstrated that the active-constrained consensus can be attained. Finally, simulation examples show the validity of the conclusions.
{"title":"Distributed Active-Constrained Consensus of Second-Order Multi-Agent Systems","authors":"Mengxiang Zeng;Peng Lin","doi":"10.1109/LCSYS.2025.3626269","DOIUrl":"https://doi.org/10.1109/LCSYS.2025.3626269","url":null,"abstract":"Most of the existing studies on position constraints integrate the constraint conditions into the agent dynamics. In contrast, the active constraints considered in this letter are implemented through control algorithms, ensuring that the agents’ states remain within their constraint sets without altering the system dynamics. Therefore, this letter explores the active-constrained consensus problem with nonconvex velocity and convex position constraints. The position constraints are fundamentally different from the velocity constraints, because the velocity constraint sets contain the origin, whereas the position constraint sets may not. The key difficulty is how to deal with the coupling between these two fundamentally different constraints. By performing a series of model transformations and utilizing the convexity of the system, it is demonstrated that the active-constrained consensus can be attained. Finally, simulation examples show the validity of the conclusions.","PeriodicalId":37235,"journal":{"name":"IEEE Control Systems Letters","volume":"9 ","pages":"2471-2476"},"PeriodicalIF":2.0,"publicationDate":"2025-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145455774","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：