-
作者：Devraj, Adithya M.; Meyn, Sean P.
作者单位：Stanford University; State University System of Florida; University of Florida
摘要：Sample complexity bounds are a common performance metric in the reinforcement learning literature. In the discounted cost, infinite horizon setting, all of the known bounds can be arbitrarily large, as the discount factor approaches unity. These results seem to imply that a very large number of samples is required to achieve an epsilon-optimal policy. The objective of the present work is to introduce a new class of algorithms that have sample complexity uniformly bounded over all discount fact...
-
作者：Krishnamoorthy, Dinesh
作者单位：Harvard University
摘要：Approximating model predictive control (MPC) policy using expert-based supervised learning techniques requires labeled training datasets sampled from the MPC policy. This is typically obtained by sampling the feasible state space and evaluating the control law by solving the numerical optimization problem offline for each sample. Although the resulting approximate policy can be cheaply evaluated online, generating large training samples to learn the MPC policy can be time-consuming and prohibi...
-
作者：Lin, Feng; Wang, Le Yi; Chen, Wen; Polis, Michael P.
作者单位：Wayne State University; Wayne State University; Oakland University
摘要：Observability of a hybrid system is defined as the ability to determine the continuous state of the system. Whether a hybrid system is observable or not depends on which events can be disabled, which events can be forced, and the connectivity of the discrete states, as well as its continuous dynamics. We model a hybrid system using a hybrid machine that takes into consideration both continuous variables and discrete events. We classify hybrid systems into four classes based on their discrete-e...
-
作者：Yan, Shuhao; Goulart, Paul J.; Cannon, Mark
作者单位：Cornell University; University of Oxford
摘要：This article considers linear discrete-time systems with additive disturbances and designs a model predictive control (MPC) law incorporating a dynamic feedback gain to minimize a quadratic cost function subject to a single chance constraint. The feedback gain is selected online, and we provide two selection methods based on minimizing upper bounds on predicted costs. The chance constraint is defined as a discounted sum of violation probabilities on an infinite horizon. By penalizing violation...
-
作者：Zheng, Lunan; Zhang, Zhijun
作者单位：South China University of Technology; South China University of Technology
摘要：By incorporating the redefined error monitor function into the network design, an error redefinition neural network (ERNN) is proposed to control mobile redundant manipulators to execute the tracking task in this article. The global asymptotic stability and the strong antidisturbance capability of the ERNN are proved theoretically. Furthermore, the ERNN can overcome the overshoot and constant disturbance. Meanwhile, the ERNN is input-to-state stable, while the bounded time-varying disturbance ...
-
作者：Cavraro, Guido; Dall'Anese, Emiliano; Comden, Joshua; Bernstein, Andrey
作者单位：United States Department of Energy (DOE); National Renewable Energy Laboratory - USA; University of Colorado System; University of Colorado Boulder
摘要：The article investigates the problem of estimating the state of a time-varying system with a linear measurement model; in particular, the article considers the case where the number of measurements available can be smaller than the number of states. In lieu of a batch linear least-squares approach-well-suited for static networks, where a sufficient number of measurements could be collected to obtain a full-rank design matrix-the article proposes an online algorithm to estimate the possibly tim...
-
作者：Cui, Yukang; Shen, Jun; Zhang, Wei; Feng, Zhiguang; Gong, Xin
作者单位：Shenzhen University; Peng Cheng Laboratory; Nanjing University of Aeronautics & Astronautics; Chongqing Three Gorges University; Harbin Engineering University; Shenzhen University; University of Hong Kong
摘要：This article studies the positivity and stability of homogeneous coupled differential-difference equations with time-varying delays. First, a sufficient positivity condition is proposed for the nonlinear coupled differential-difference equations with delays. Then, based on this positivity condition, we present necessary and sufficient conditions ensuring the exponential stability and bounding the decay rate for time-delay homogeneous coupled differential-difference equations with homogeneity o...
-
作者：Kasis, Andreas; Lestas, Ioannis
作者单位：University of Cyprus; University of Cyprus; University of Cyprus; University of Cambridge
摘要：Thermostatically controlled loads (TCLs) can provide ancillary services to the power network by aiding existing frequency-control mechanisms. TCLs are, however, characterized by an intrinsic limit cycle behavior, which raises the risk that these could synchronize when coupled with the frequency dynamics of the power grid, i.e., simultaneously switch, inducing persistent and possibly catastrophic power oscillations. To address this problem, schemes with a randomized response time in their contr...
-
作者：Lindemann, Lars; Pappas, George J.; Dimarogonas, Dimos, V
作者单位：University of Pennsylvania; Royal Institute of Technology
摘要：The deployment of autonomous systems in uncertain and dynamic environments has raised fundamental questions. Addressing these is pivotal to build fully autonomous systems and requires a systematic integration of planning and control. We first propose reactive risk signal interval temporal logic (ReRiSITL) as an extension of signal temporal logic (STL) to formulate complex spatiotemporal specifications. Unlike STL, ReRiSITL allows to consider uncontrollable propositions that may model humans as...
-
作者：Ren, Wei; Dimarogonas, Dimos, V
作者单位：Universite Catholique Louvain; Royal Institute of Technology
摘要：This article studies the tracking control problem of networked multiagent systems under both multiple networks and event-triggered mechanisms. Multiple networks are to connect multiple agents and reference systems with decentralized controllers to guarantee their information transmission, whereas the event-triggered mechanisms are to reduce the information transmission via the networks. In this article, each agent has a network to communicate with its controller and reference system, and all n...
