Stochastic learning and optimization—A sensitivity-based approach

We introduce a sensitivity-based view to the area of learning and optimization of stochastic dynamic systems. We show that this sensitivity-based view provides a unified framework for many different disciplines in this area, including perturbation analysis, Markov decision processes, reinforcement learning, identification and adaptive control, and singular stochastic control; and that this unified framework applies to both the discrete event dynamic systems and continuous-time continuous-state systems. Many results in these disciplines can be simply derived and intuitively explained by using two performance sensitivity formulas. In addition, we show that this sensitivity-based view leads to new results and opens up new directions for future research. For example, the n th bias optimality of Markov processes has been established and the event-based optimization may be developed; this approach has computational and other advantages over the state-based approaches.     

A two-degree-of-freedom PI controller based on events

In this paper, a completely event-based two-degree-of-freedom proportional-integral controller is presented. The architecture of the controller is based on event-based decoupled solutions for the set-point following and the load disturbance rejection tasks. For the first task, the solution is a design procedure that, by considering a first-order-plus-dead-time model of the process and a predesigned open-loop control action, produces an event-based feedforward controller that provides the required process variable transition with just two events. The second task is solved by applying two separate event-based proportional and integral control actions. Because the two tasks are initially solved independently, two solutions for coupling them are described. Illustrative examples of the performance of the controller are included as well as experimental results.     

一种Internet计算中的事件观察与通知机制设计框架

提出了一个事件机制设计的框架,定义了事件的各个方面特性。这个框架将事件用7个模型来描述：对象模型、事件模型、命名模型、观察模型、时间模型、通知模型、资源模型。对每个模型进行了详细的讨论。并讨论了与Internet环境中事件机制相关的工作,提出了可能出现的研究热点问题。     

EVA: An event algebra supporting complex event specification

In applications such as digital libraries, stock tickers, traffic control, or supply chain management, composite events have been introduced to enable the capturing of rich situations. Composite events seem to follow common semantics. However, on closer inspection we observed that the evaluation semantics of events differs substantially from system to system. In this paper, we propose a parameterized event algebra that defines the detailed semantics of composite event operators. We introduce composite event operators that support explicit parameters for event selection and event consumption. These parameters define how to handle duplicates in both primitive and composite events.The event algebra EVA forms the foundation for a unifying reference language that allows for translation between arbitrary event composition languages using transformation rules. This translation, in turn, enables a mediator service that can federate heterogeneous event-based systems. Our approach supports the seamless integration of event-based applications and changing event sources without the need to redefine user profiles. The event algebra is exemplified in the domain of transportation logistics.     

基于事件的发布-订阅系统模型

Internet的广泛应用已经改变了分布式系统的规模,使得传统的基于请求/应答的点对．点的同步通信已不能很好地满足大规模的动态分布式应用环境。为了加强大规模的分布式环境中实体之间的通信协作,系统要求更加灵活的通信模型,以反映应用的动态和非耦合特性。基于事件通信的中间件是建立大规模分布式系统的有效方式,发布者/订阅者（Publisher/subscriber）是目前广泛使用的基于事件的通信模型,支持发布者和订阅者之间在时间、空间和同步方面的非耦合以及多对多的通信模式,提供大规模系统所要求的交互间的松散耦合。本文详细阐述了事件系统的组成以及基于事件系统的PUB／SUB模型,提出了基于PUB／SUB的事件系统中间件的核心及实现的关键,这些中间件为大规模的分布式应用提供更多和更有力的保障。     

Scheduler-based state estimation over multiple channels networks

We investigate the remote state estimation problem for networked systems over parallel noise-free communication channels. Due to limited network capabilities in practical network environments, communication schedulers are implemented at the transmit side of each subchannel to promote resource efficiency. Specifically, the processed signals are transmitted only when it is necessary to provide the real-time measurements to the remote estimator. The recursive approximate minimum mean-square error estimator is established to restore the state vector of a target plant by utilizing the scheduled transmission signals. All the information coming from the individual subchannels, even if no measurement is sent, will contribute to improve the estimation performance in an analytical form. Finally, a numerical example is given to illustrate the effectiveness of the main results.     

一种多时相遥感影像存储管理的新方法

针对海量遥感影像时间分辨率不断提高,多时相遥感影像存储空间耗费大、查询效率低的问题,提出一种针对多时相遥感影像数据存储管理的模型和方法。首先确定多时相遥感影像的基态影像,然后运用变化检测技术计算出不同时相间影像的差异,将差异变化值大于设定阈值的遥感影像区域进行存储,并用于对基态影像的修正。实验证明,运用该方法管理多时相遥感影像时,可有效节省存储空间。     

Decentralized event-triggered consensus with general linear dynamics

The consensus problem with general linear dynamics and undirected graphs is studied in this paper by means of event-triggered control strategies. A novel consensus protocol is proposed, where each agent implements a model of the decoupled dynamics of its neighbors. Under this control strategy, transmission of information does not occur continuously but only at discrete points in time. The approach presented in this paper provides both a decentralized control law and a decentralized communication policy. We are able to design thresholds that only depend on local information and guarantee asymptotic consensus. Positive inter-event times are guaranteed for particular cases of the linear dynamics. In an extension, a positive constant is added to the thresholds in order to exclude Zeno behavior for general linear dynamics. The difference between agents trajectories can be bounded in this case and bounds on the state disagreement are derived.     

Extension and experimental evaluation of an event-based state-feedback approach

Event-based control aims at reducing the traffic load of the communication network used to implement feedback control. The communication is invoked only if an event indicates that the control error exceeds a certain threshold. This paper extends a state-feedback approach to event-based control published recently and proves that the extended control loop asymptotically reaches the set-point for arbitrary constant disturbances. This control method is applied to a thermofluid process to illustrate the main characteristics of event-based control and to evaluate the closed-loop performance. Experiments with a pilot plant show that the event-based control scheme is robust against severe model uncertainties.