Abstract:In a distributed virtual environment (DVE), the bottleneck of performance is the communication among hosts which keeps the consistency of virtual entities. The neural-network models of state vectors for unpredictable entities in a virtual scene are build in this paper, and the functional-link net to real-time prediction of their behavior is applied. Firstly, the principles and characters of functional-link net are introduced in this paper; secondly, after description of the traditional dead reckoning (DR) algorithm, an adaptive version of the algorithm based on functional-link net is presented; the network software architecture based on the algorithm is also given; finally, an example of the algorithm is given with experimental data, which shows the good performance of it.

Abstract:A trace-based logic language: L_Trace is defined in this paper, which is an extension of the first-order linear temporal logic and serves as cornerstone of the works ——Object Calculus. The objects in trace calculus represent the dynamic entities endowed with a local state and external actions, and described by an object signature in syntax. An object signature is a 4-tuple Ω=(S,F,A,E) in which S stands for a set of data sorts, F functions, A attributes and E actions. Σ=(S,F) is nothing but a usual signature in the context of algebraic specification. It can be extended to ΣE with the action regarded as a special data sort. The semantics of trace calculus is defined by an object signature semantic interpreting structure I=(A,F,E), which consists of a ΣE- Algebra A giving an interpretation about data parameters, a mapping F evaluating the attributes on an action trace, and a relation E giving a relationship between actions and a trace. Finally, we contribute an axiom system of our trace calculus and outlines its proof of soundness after we define the syntax and semantics of the trace calculus.

Abstract:Cooperative caching is one of the key technologies used to improve the performance of a cluster file system. s2fs (scalable single-image file system), a single-image cluster file system prototype, uses dual-granularity cache coherence protocol in order to implement efficient cooperative caching which meets the needs for strict UNIX-semantics. In this paper, a hint-based heuristic cache replacement algorithm under s2fs' dual-granularity protocol is proposed, and the analytical models are established for heuristic algorithm and one of the existing coordinated algorithm——N-chance. The analytical results show that the heuristic algorithm can effectively reduce the I/O response time compared with N-chance algorithm almost in each case.

Abstract:MFS is a program development system based on rewriting techniques. The language provided by MFS, which called MFSL, is a mixed language that combines enhanced functional language and algebraic speci-fication language. Optimal rules in MFSL can improve efficiency and satisfy specific requirements. Both the type system in MFSL and the mechanism of proving and testing in MFS can help the programmers to find problems early and can improve the correctness of program. The efficiency of the implementation of the system developed by MFS is high due to the techniques used by MFS such as needed parallel outermost reduction strategy,graph reduction and witnessed test set approach.Higher efficiency and correnctness of program can be developed by MFS in a shorter period.

Abstract:A warehouse is a data repository containing integrated information for efficient querying and analysis, which data come from the databases or the other info-sources at different places. Materialized view is the primary information entity stored in the data warehouse. It must be refreshed when the corresponding data changed in the database. Consistency problem will be arisen during materialized views on-line maintenance because of the invention of OLAP query. In this paper, the authors introduce a new algorithm, MVCA (multiversion compensating algorithm), using multiversion and compensating techniques, along with acknowledgement mechanism to synchronize the maintenance process between the database and the data warehouse so as to ensure the data consistency. At the end of the paper, the authors illustrate the application of the algorithm by a typical example.

Abstract:To overcome the problem of slow convergence before the genetic algorithms (GAs) reach the global optima, it is an effective way to combine the conventional local search algorithms with GAs. A new hybrid algorithm that incorporates the flexible polyhedron method into the orthogonal genetic algorithm (OGA) is presented in this paper. The experiments showed that it can achieve better performance by performing global search and local search alternately. The new algorithm can be applied to solve the function optimization problems efficiently.

Abstract:The online bibliographic retrieval service (OBRS) is a main service which should be provided by libraries on network environment. Every OBRS provided via Telnet or Web has its own access procedures and a special user interface and query language. The Z39.50 protocol offers a solution to heterogeneity problem. In this paper, a system that provides OBRS based on Z39.50 is described. The system supports Chinese retrieval and multi-MARC (machine-readable cataloguing) including CNMARC, and is highly configurable and scaleable. Main technical design and simple comparison with related systems are given in this paper.

Abstract:The social commitment between agents needs to be built for problem solving in the multi-agent systems. A mechanism of social commitment by combining Belief-Desire-Intention and Utility theory is presented in this paper, which guarantees agent's rational action on logic and decision. The Rao & Georgeff's Belief-Desire-Intention theory and S.Kraus's idea of incentive contract are combined ,the Haddadi's mechanism of social commitment is modified, and the Castelfranchi's concept of social commitment with implementing support is provided in this paper.

Abstract:This paper concentrates on the important techniques of design and implementation of Interface Agent (NanDa User Interface Agent, NDUIA) in DOLTRI-Agent (distance and open learning training resource information retrieval agent). This system extends the memory-based reasoning technique, using multiple memory models and multiple analysis models, and the authors can generate a specific interest's model for each user by analyzing his/her using experience. At the same time, through the combination of user interests model and using experience in special scene, interface agent can provide some active intelligent services, such as information guidance, preprocess of the searching results, automatic modification of classification information, intelligent and instant help, etc. In this way, the authors achieve collaboration between software and its users.

Abstract:In this paper, the authors suggest to think about the problem of process metric from the process modeling point of view, and discuss a methodology for metric-supported software process modeling, which includes a goal-oriented process metric modeling method, a descriptive mechanism for software process and an algorithm for metric-supported software process modeling.

Abstract:Grammatical Inference (GI) is a problem of inductive learning of formal languages, which deals with how to obtain the grammatical description of a formal language from given finite data drawn from the language. In this paper, the authors provide a survey of the history and recent advances in GI field. They first present some learning models for GI. Then, they enumerate methods for GI with an emphasis on the results concerning the inference of context free grammar class and its some subclasses, hidden Markov models, and stochastic context-free grammar class. At last, they briefly give some applications of GI as well as the future directions of GI research.

Abstract:The complexity of developing parallel software makes it attractive to research their reusability. The distributed object technology provides some benefits in this field because it can encapsulate parallel programs to components and exploit the parallelism in the various heterogeneous systems. However, the drawback to use this technology to do parallel computing is obvious because of the limit interoperability among those components. A model of interoperability is proposed in this paper. This model keeps compliant with the common distributed object model while gaining more parallelism from the interoperability of distributed objects.

Abstract:In this paper, the shortcomings in theory and limitation in applications of FOIL are analysed. To overcome these difficulties, instance graph H_(R,E) and instance order are introduced to clarify the relationship between the set R of recursive rules and the instance space E. Based on these concepts, a new ILP algorithm, FOILPlus, is put forward, which prevents the generation of harmful recursive rules by utilizing hung example and hung arc to hold Instance Graph. The algorithm can complete learning tasks without the restriction of constant ordering, and does not substantially raise the computational complexity compared with FOIL. FOILPlus has been implemented, and experiments with it show that it does complete two learning tasks which FOIL fails.

Abstract:In this paper, the authors discuss IS-Global's pivotal technique in design and implementation. IS-Global is a relational multi-database system, which provides an integrated global information system platform for enterprises applying CIMS (computer integrated manufacture system), and hides the detail of the heterogeneity of network, local operating systems and database systems. The major characteristics of IS-Global are as follows: (1) It is a tightly integrated system with unique global schema. (2) Being a multi-user system, IS-Global possesses two-lever client/server architecture and supports concurrent and security control mechanisms. (3) IS-Global provides graph-based tools, ODBC function call and interactive SQL interface. (4) IS-Global supports multi-location queries.

Abstract:Semistructured data has irregular or incomplete structure. In recent research on semistructured data sources and integration for heterogeneous data sources, models for semistructured data are based on direct graph with root vertex, so querying semistructured data is equivalent with searching in graph. In addition, path with wildcard characters brings more complexity in query processing. In this paper, the authors present the strategies deployed in querying and optimizing OIM (model for object integrating) data in Versatile——a system for integrating heterogeneous data sources. Algorithms for generating query plan and extending path are discussed in detail and three optimization methods, path index (Pindex), level index(Lvindex) and knowledge of data source are introduced. Also the approach can be applicable to other graph-based semistructured data easily.

Abstract:Multicast capability at the network layer is necessary in order to reduce the bandwidth requirements of multiparty, multicast applications. Integral to such a capability are multicast address management and multicast routing mechanisms. Multicast address management and connection control are two essential components needing to be solved in the Multipoint applications. Based on hierarchical autonomous structure in accordance with the self-organization topologies of INTERNET, this paper puts forward a multicast address management scheme which sets up the hierarchical partition, which is dynamic, of Multicast address space in order to reduce the blocking probability. Only MAM (multicast address management) in the bottom-layer domain may allocate directly the addresses to multipoint session so that the session acquire the address more rapidly in average. And, a connection control hierarchy (CCH) based on master/slave relationship and a simple efficient building algorithm of multi-point connection are also advanced. The paper also describes the normal operations of multicast address management and multi-point connections controller. With simulation experiment, three of multicast address allocation modes, hierarchical autonomous mode (HAM), central mode (CM) and network-divided mode (NDM) (Eleftheriadis proposed), are compared. The result shows that hierarchical autonomous mode integrates the merits of central mode and network-divided mode, owns high efficiency in the whole. Connection control hierarchy also is efficient in the connection control. They are shown to be highly robust, flexible and scalable.