Image Annotation by Graph-Based Inference With Integrated Multiple/Single Instance Representations

In most of the learning-based image annotation approaches, images are represented using multiple-instance (local) or single-instance (global) features. Their performances, however, are mixed as for certain concepts, the single-instance representations of images are more suitable, while for others, the multiple-instance representations are better. Thus this paper explores a unified learning framework that combines the multiple-instance and single-instance representations for image annotation. More specifically, we propose an integrated graph-based semi-supervised learning framework to utilize these two types of representations simultaneously. We further explore three strategies to convert from multiple-instance representation into a single-instance one. Experiments conducted on the COREL image dataset demonstrate the effectiveness and efficiency of the proposed integrated framework and the conversion strategies.      

Language equations for timed alternating finite automata

Traditionally, finite state automata are untimed or asynchronous models of computation in which only the ordering of events, not the time at which events occur, would affect the result of a computation. For real-time systems, it is important to augment these models of computation with a notion of time. For this purpose timed automata have become a powerful canonical model for describing timed behaviors and an effective tool for modeling real-time computations. In this paper, we extend the notion of timed alternating finite automata (TAFA), a class of alternating finite automata (AFA) extended with a finite set of real-valued clocks, and we present an algebraic interpretation of TAFA which parallels that of timed regular expressions and language equations. We further extend the equational representation of AFA to describe timed alternating finite automata, and explore solutions for such equations over time languages.     

Constraint-based modeling of discrete event dynamic systems

Numerous frameworks dedicated to the modeling of discrete event dynamic systems have been proposed to deal with programming, simulation, validation, situation tracking, or decision tasks: automata, Petri nets, Markov chains, synchronous languages, temporal logics, event and situation calculi, STRIPS…All these frameworks present significant similarities, but none offers the flexibility of more generic frameworks such as logic or constraints. In this article, we propose a generic constraint-based framework for the modeling of discrete event dynamic systems, whose basic components are state, event, and time attributes, as well as constraints on these attributes, and which we refer to as CNT for Constraint Network on Timelines. The main strength of such a framework is that it allows any kind of constraint to be defined on state, event, and time attributes. Moreover, its great flexibility allows it to subsume existing apparently different frameworks such as automata, timed automata, Petri nets, and classical frameworks used in planning and scheduling.     

Accepting Zeno words: a way toward timed refinements

Timed models were introduced to describe the behaviors of real-time systems and they were usually required to produce only executions with divergent sequences of times. However, when some physical phenomena are represented by convergent executions, Zeno words appear in a natural way. Moreover, time can progress if such an infinite execution can be followed by other ones. Therefore, in a first part, we extend the definition of timed automata, allowing to generate sequences of infinite convergent executions, while keeping good properties for the verification of systems: emptiness is still decidable. In a second part, we define a new notion of refinement for timed systems, in which actions are replaced by recognizable Zeno (timed) languages. We study the properties of these timed refinements and we prove that the class of transfinite timed languages is the closure of the usual one (languages accepted by Muller or Büchi timed automata) under refinement. Received: 16 October 1998 / 8 March 2000     

基于Uppaal的实时系统AADL数据流模型的转换与验证

体系结构分析设计语言AADL是一种可支持软硬件一体化建模及同一模型多元分析的形式化与图形化建模语言。采用时间自动机形式化模型检验方法对AADL模型中的数据流进行转换和验证。考虑到单一数据流与混合数据流的差异性,分别设计了数据流到时间自动机模型的转换规则,并通过时间自动机网络实现数据流的综合分析。设计开发了自动化模型转换的插件AADLToUppaal Plug-in,将其嵌入到OSTATE工具中,使用时间自动机建模与验证工具Uppaal对转换得到的时间自动机进行模拟和验证,等价地验证所设计的AADL模型数据流时延是否满足系统实时性要求。仿真实验结果表明,所设计的数据流模型转换方法能有效地将AADL模型转换到时间自动机模型,并能在Uppaal中正确地分析原模型的数据流时延特性。     

An algebraic approach to data languages and timed languages

Algebra offers an elegant and powerful approach to understand regular languages and finite automata. Such framework has been notoriously lacking for timed languages and timed automata. We introduce the notion of monoid recognizability for data languages, which includes timed languages as special case, in a way that respects the spirit of the classical situation. We study closure properties and hierarchies in this model and prove that emptiness is decidable under natural hypotheses. Our class of recognizable languages properly includes many families of deterministic timed languages that have been proposed until now, and the same holds for non-deterministic versions.     

Controller synthesis for dynamic hierarchical real-time plants using timed automata

We use timed I/O automata based timed games to synthesize task-level reconfiguration services for cost-effective fault tolerance in a case study. The case study shows that state-space explosion is a severe problem for timed games. By applying suitable abstractions, we dramatically improve the scalability. However, timed I/O automata do not facilitate algorithmic abstraction generation techniques. The case study motivates the development of timed process automata to improve modeling and analysis for controller synthesis of time-critical plants which can be hierarchical and dynamic. The model offers two essential features for industrial systems: (i) compositional modeling with reusable designs for different contexts, and (ii) state-space reduction technique. Timed process automata model dynamic networks of continuous-time communicating plant processes which can activate other plant processes. We show how to establish safety and reachability properties of timed process automata by reduction to solving timed games. To mitigate the state-space explosion problem, an algorithmic state-space reduction technique using compositional reasoning and aggressive abstractions is also proposed. In this article, we demonstrate the theoretical framework of timed process automata and the effectiveness of the proposed state-space reduction technique by extending the case study.     

带时间和资源约束的实例化过程模型验证方法

软件过程的性能是由软件过程模型和软件过程实例化两方面因素决定,如果对软件过程进行了不恰当的实例化,会导致成本超支、进度延期、甚至项目失败．已有的过程描述法不足以分析实例化过程模型,由于没有考虑实例化阶段的时间资源约束,语法结构正确的过程模型并不能保证过程执行的正确性．提出一种带时间和资源约束的实例化过程模型验证方法,为目前已有的s-TRISO／ML建模语言增加时间和资源约束属性,然后提出了从s-TRISO／ML模型转换成时间自动机的转换方法和实现算法,利用已有的分析工具Uppaal对转换得到的时间自动机的性质进行验证,得到一个合理的实例化模型,从而为真实的开发流程提供指导．     

Model-checking timed automata with deadlines with Uppaal

Timed automata with deadlines (TAD) are a form of timed automata that admit a more natural representation of urgent actions, with the additional advantage of avoiding the most common form of timelocks. We offer a compositional translation of a practically useful subset of TAD to timed safety automata (the well-known variant of timed automata where time progress conditions are expressed by invariants). More precisely, we translate networks of TAD to the modeling language of Uppaal, a state-of-the-art verification tool for timed automata. We also describe an implementation of this translation, which allows Uppaal to aid the design and analysis of TAD models.     

<Emphasis Type="Italic">SetExp</Emphasis>: a method of transformation of timed automata into finite state automata

Real-time discrete event systems are discrete event systems with timing constraints, and can be modeled by timed automata. The latter are convenient for modeling real-time discrete event systems. However, due to their infinite state space, timed automata are not suitable for studying real-time discrete event systems. On the other hand, finite state automata, as the name suggests, are convenient for modeling and studying non-real time discrete event systems. To take into account the advantages of finite state automata, an approach for studying real-time discrete event systems is to transform, by abstraction, the timed automata modeling them into finite state automata which describe the same behaviors. Then, studies are performed on the finite state automata model by adapting methods designed for non real-time discrete event systems. In this paper, we present a method for transforming timed automata into special finite state automata called Set-Exp automata. The method, called SetExp, models the passing of time as real events in two types: Set events which correspond to resets with programming of clocks, and Exp events which correspond to the expiration of clocks. These events allow to express the timing constraints as events order constraints. SetExp limits the state space explosion problem in comparison to other transformation methods of timed automata, notably when the magnitude of the constants used to express the timing constraints are high. Moreover, SetExp is suitable, for example, in supervisory control and conformance testing of real-time discrete event systems.     

Timed substitutions for regular signal-event languages

In the classical framework of formal languages, a refinement operation is modeled by a substitution and an abstraction by an inverse substitution. These mechanisms have been widely studied, because they describe a change in the specification level, from an abstract view to a more concrete one, or conversely. For timed systems, there is up to now no uniform notion of substitution. In this paper, we study timed substitutions in the general framework of signal-event languages, where both signals and events are taken into account. We prove that regular signal-event languages are closed under substitution and inverse substitution. To obtain these results, we use in a crucial way a “well known” result: regular signal-event languages are closed under intersection. In fact, while this result is indeed easy for languages defined by Alur and Dill’s timed automata, it turns out that the construction is much more tricky when considering the most involved model of signal-event automata. We give here a construction working on finite and infinite signal-event words and taking into account signal stuttering, unobservability of zero-duration τ-signals and Zeno runs. Note that if several constructions have been proposed in particular cases, it is the first time that a general construction is provided.     

CPO semantics of timed interactive actor networks

We give a denotational framework for composing interactive components into closed or open systems and show how to adapt classical domain-theoretic approaches to open systems and to timed systems. For timed systems, prior approaches are based on temporal logics, automata theory, or metric spaces. In this paper, we base the semantics on a CPO with a prefix order, as has been done previously for untimed systems. We show that existence and uniqueness of behaviors are ensured by continuity with respect to this prefix order. Existence and uniqueness of behaviors, however, do not imply that a composition of components yields a useful behavior. The unique behavior could be empty or smaller than expected. We define liveness and show that appropriately defined causality conditions ensure liveness and freedom from Zeno conditions. In our formulation, causality does not require a metric and can embrace a wide variety of models of time.     

Modeling and analysis of timed Petri nets using heaps of pieces

The authors show that safe timed Petri nets can be represented by special automata over the (max, +) semiring, which compute the height of heaps of pieces. This extends to the timed case the classical representation a la Mazurkiewicz of the behavior of safe Petri nets by trace monoids and trace languages. For a subclass including all safe free-choice Petri nets, we obtain reduced heap realizations using structural properties of the net (covering by safe state machine components). The authors illustrate the heap-based modeling by the typical case of safe jobshops. For a periodic schedule, the authors obtain a heap-based throughput formula, which is simpler to compute than its traditional timed event graph version, particularly if one is interested in the successive evaluation of a large number of possible schedules     

Determinization of timed Petri nets behaviors

In this paper we are interested in sequentialization of formal power series with coefficients in the semiring \((\mathbb {R}\cup \{- \infty \},\max ,+)\) which represent the behavior of timed Petri nets. Several approaches make it possible to derive nondeterministic (max, + ) automata modeling safe timed Petri nets. Their nondeterminism is a serious drawback since determinism is a crucial property for numerous results on (max, + ) automata (in particular, for applications to performance evaluation and control) and existing procedures for determinization succeed only for restrictive classes of (max, + ) automata. We present a natural semi-algorithm for determinization of behaviors based on the semantics of bounded timed Petri nets. The resulting deterministic (max, + ) automata can be infinite, but a sufficient condition called strong liveness is proposed to ensure the termination of the semi-algorithm. It is shown that strong liveness is closely related to bounded fairness, which has been widely studied for Petri nets and other models for concurrency. Moreover, if the net cannot be sequentialized we propose a restriction of its logical behavior so that the sufficient condition becomes satisfied for the restricted net. The restriction is based on the synchronous product with non injectively labeled scheduler nets that are built in an incremental hierarchical way from simple scheduler nets.     

Timed automata with urgent transitions

We propose an extension to the formalism of timed automata by allowing urgent transitions. An urgent transition is a transition which must be taken within a fixed time interval from its enabling time and it has higher priority than other non-urgent transitions enabled in the same state. We give a set of rules formally describing the behavior of urgent transitions and we show that, from a language theoretic point of view, the addition of urgency does not improve the expressive power of timed automata. From a specification point of view, the use of urgent transitions allows shorter and clear specifications of behaviors involving urgency and priority. We use timed automata with urgent transitions for specifying a multicast protocol for mobile computing.Received: 10 January 2003, Published online: 17 February 2004A first version of this paper appeared in [15]     

Forward Analysis of Updatable Timed Automata

Timed automata are a widely studied model. Its decidability has been proved using the so-called region automaton construction. This construction provides a correct abstraction for the behaviours of timed automata, but it suffers from a state explosion and is thus not used in practice. Instead, algorithms based on the notion of zones are implemented using adapted data structures like DBMs. When we focus on forward analysis algorithms, the exact computation of all the successors of the initial configurations does not always terminate. Thus, some abstractions are often used to ensure termination, among which, a widening operator on zones.In this paper, we study in detail this widening operator and the corresponding forward analysis algorithm. This algorithm is most used and implemented in tools like KRONOS and UPPAAL. One of our main results is that it is hopeless to find a forward analysis algorithm for general timed automata, that uses such a widening operator, and which is correct. This goes really against what one could think. We then study in detail this algorithm in the more general framework of updatable timed automata, a model which has been introduced as a natural syntactic extension of classical timed automata. We describe subclasses of this model for which a correct widening operator can be found.     

Scheduler Modeling Based on the Controller Synthesis Paradigm

The controller synthesis paradigm provides a general framework for scheduling real-time applications. Schedulers can be considered as controllers of the applications; they restrict their behavior so that given scheduling requirements are met. We study a modeling methodology based on the controller synthesis paradigm. The methodology allows to get a correctly scheduled system from timed models of its processes in an incremental manner, by application of composability results which simplify schedulability analysis. It consists in restricting successively the system to be scheduled by application of constraints defined from scheduling requirements. The latter are a conjunction of schedulability requirements that express timing properties of the processes and policy requirements about resource management. The presented methodology allows a unified view of scheduling theory and approaches based on timing analysis of models of real-time applications.     

离散时段演算的符号模型验证

模型验证是对有限状态系统的一种形式化确认方法，近几年，模型验证方法已逐步扩展到实时系统应用中，为解决实时系统的模型验证问题，本文采用离散时段演算人实时系统规格说明的形式语言，用时间自动机作为实时系统的实现模型，对模型验证问题进行了细致的分析，并提出了一种具有实际应用价值的方法－商技术，该方法可以在避免当多个时间自动机并行组合时可能产生的状态空间组合爆炸问题，同时还可以简化整个模型验证问题。     

A Kleene–Schützenberger theorem for weighted timed automata

During the last years, weighted timed automata have received much interest in the real-time community. Weighted timed automata form an extension of timed automata and allow us to assign weights (costs) to both locations and edges. This model, introduced by Alur et al. (2001) and Behrmann et al. (2001), permits the treatment of continuous consumption of resources and has led to much research on scheduling problems, optimal reachability and model checking. Also, several authors have derived Kleene-type characterizations of (unweighted) timed automata and their accepted timed languages. The goal of this paper is to provide a characterization of the behaviours of weighted timed automata by rational power series. We define weighted timed automata with weights taken in an arbitrary semiring, resulting in a model that subsumes several weighted timed automata concepts of the literature. For our main result, we combine the methods of Schützenberger, a recent approach for a Kleene-type theorem for unweighted timed automata by Bouyer and Petit as well as new techniques. Our main result also implies Kleene-type theorems for several subclasses of weighted timed automata investigated before, e.g., for timed automata and timed automata with stopwatch observers.