Catálogo de publicaciones - libros

Synchronization is used in dual redundant fault-tolerant system to make two computers work jointly. It determines the work mode and controls the operations of the system. The paper presents a dual redundant fault-tolerant system and proposes its process of synchronization based on task. The synchronization treats task as the minimal operation unit. And it is implemented with the assist of dual-computer-controller and outer memory, the latter includes task buffer and global data region. Dual-computer-controller controls the input and output of tasks that are stored in task buffer. The switch in of backup computer is implemented by duplicating global data from global data region used by host computer. Additional resolutions for key points, such as the switch between two computers, are also put forward in conclusion. It is proved that the task-level synchronization can make two computers work in phase, and the system is applicable to critical fields requiring high dependability.

This paper describes application of association rules in education. To make everything more clearly visible a graphic display of objects and attributes in a lattice structure is provided. Vectors and matrices are used to reduce the computational complexity while searching for association rules in the case when at least one attribute included in the ‘if’ part of the statement is known.

This paper briefly introduces the main ideas of Machine Translation (MT) techniques, then discusses the application of MT in the China-America Digital Academic Library (CADAL)

The active functions of aDBS (active Database System) are currently achieved by ECA rule system. But it is difficult for EC A rules system to be analyzed and validated. And there is no perfect model today. Based on the Petri Net, This paper thinks over the character of rules system, and establishes a system model. And some behaviors are analyzed in this paper.

To reason the interaction between agents’ epistemic states and environments, the agents’ epistemic states should include ontologies of agents (the agents’ assumptions about environments), what the agents know, believe and desire, and also should include the agents’ interpretations about these symbols in ontologies onto the environments. A formal description of such agents’ epistemic states and several examples are given to show how such a description can be used to explain the puzzles in reasoning the modal sentences. For example, a sentence holding in a structure may have different consequences reasoned by different agents, because of the different epistemic states.

Security of networking systems has been an issue since computer networks became prevalent, most especially now that Internet is changing the facie computing. Intrusions pose significant threats to the integrity, confidentiality and availability of information for the internet users. In this paper, a new approach to real-time network anomaly intrusion detection via Fuzzy-Bayesian is proposed to detect malicious activity against computer network; the framework is described to demonstrate the effectiveness of the technique. The combination of fuzzy with Bayesian classifier will improve the overall performance of Bayes based intrusion detection system (IDS). Also, the feasibility of our method is demonstrated by the experiment performed on KDD 1999 IDS data set.

This paper introduces the reference theory and algorithm of text categorization by using fuzzy cognitive map(FCM), which is based on value inference and can be able to infer by combing rule and statistics. This method is flexible and robust, and we do not need train the corpus time after time, it is suitable to the text categorization of insufficiency training, new subject and multi-classification.

The Semantic Web is charming but not easy to realize, since one of the prerequisite steps is to create semantic and precise data adhere to traditional web pages. We provide an annotation system ConAnnotator, which is an ontology-based annotation system and allows cooperative working. It aims at supporting the annotation process as well as the evolvement of the ontology. By semi-automatically creating ontology-based annotations and managing the statistic information about the annotation history, it facilitates the annotation process, makes the annotated documents connected to the ontology and further constructs the Semantic Web, and ultimately helps the users to evolve the ontology itself.

This paper reports a method of simulate an Agent emotion in a Virtual Super Market environment. Utility Function is the key to depict people’s emotion especially in the satisfaction feelings. This method finished the process from the emotion to the decision. Agent will act emotionally according to the Utility.

Data mining implies “digging through tons of data” to uncover implicit information. Software defect is an essential characteristic of software development process, offering much information about software quality assurance. Based on the information supplied by defects, it can help a software team with capability of software quality assurance. In order to dig through lots of defects to uncover implicit information of defects, this paper integrates PCA and DA, combined with the relationship among each defect attribute index, and then we can dig critical factor out of software defects. These critical factors are the vital few defects affecting software quality, warning programmer to stress concentration on getting rid of these vital few defects.