Abstract:A fast and efficient triangulation algorithm is presented in this paper. The scattered data on any 2D shape is triangulated by this algorithm. Triangular mesh optimization is done at the same time with mesh creating. The resulted mesh is fit to Delaunay criteria. The algorithm complexity is about linear to the number of points. Application in oil and geology exploration, the algorithm can deal with large scale scattered data including complicated faults.

Abstract:In the present paper, it is proved that the interval wavelets neural networks has universal and L² approximation properties and is a consistent function estimator. Convergence rates associated with these properties do not decrease as d increases in d-dimensional function learning, i.e., the “curse of dimensionality” is eliminated substantially. In the experiments, the proposed interval wavelet neural networks, compared to traditional wavelet networks, has performed better.

Abstract:In this paper, the authors deal with the generation of permutations and involutions with dependence degree >1. For permutation, they first give a method of generating a kind of permutations with the same dependence degree by affine transformations to a given permutation, then a recursive method of generating higher dimensional permutations from lower ones, and numbers of permutations generated by these methods are evaluated. For involution, the authors give a method of generating involutions with dependence degree >1 by making p disjoint transpositions with the same distance from fixed points of a given involution, and a loose lower bound of numbers of such generated involutions.

Abstract:Data declustering methods of parallel databases have significant effect on the performance of the parallel algorithms on the databases. If the characteristics of the declustering methods of the parallel database can be adequately exploited in designing parallel algorithms for implementing the database operations in parallel database systems, high performance parallel algorithms could be resulted. Unfortunately, the advantages of the underlying declustering methods of the parallel databases are rarely considered in existent parallel algorithms on the parallel databases. The focus of this paper is on how to utilize the advantages of the underlying declustering methods of the databases in designing parallel join algorithms. A set of parallel join algorithms based on the CMD declustering method are proposed. Theoretical and experimental results show that the proposed algorithms are more efficient than other parallel join algorithms.

Abstract:Free-form deformation through the control of parametric surfaces is a flexible technique. In the old method objects to be deformed are embedded into the parametric space of the base surface directly and the undesired distortions are frequently introduced into the result. A homogeneous deformation method is proposed in this paper which successfully eliminates or alleviates the undesired distortion. The main difference between the proposed method and the existing methods is that the new one embeds objects into the parametric space of the base surface by isometric or locally isometric mapping using the planar development of the base surface as the intermediate medium, thus the object to be deformed can follow the shape of the base surface homogeneously. Implementation results indicate the potentials of the new approach.

Abstract:The availability of server is a key factor in improving client/server system performance. The most promising approach in this regard is to use multiple replicated servers. However, the blocking problem is a bottleneck in replicated data update for this approach. In this paper, a non-blocking independently replicated data update method and a reconciliation mechanism to reach agreement at those servers not committed because of the failure are presented.

Abstract:In order to reduce the complexity in the calculation model of Nilsson's probabilistic logic, the authors presented a new approach with the super-tree-like decomposition model in this paper, and proved the completeness theorem of the given method.

Abstract:In this paper, an intelligent database tool FoxPert based on popular database systems is presented. This tool implements the seamless connection of FoxPro, Oracle and an expert programming language CLIPS (C language integrated production system) via an enhanced loose-coupling architecture. It is distinct from other available tools in the integrated platform, integrating direction and user interface. It has been used in several departments at present.

Abstract:In this paper, the authors introduce the job selection strategy in the load balancing system——ILBOT (intelligent load balancer based on on-line tracing). In this strategy, the authors consider the influence upon response time of various kind of job caused by the various load environment. According to it, they estimate the benefit and overhead of transferring a job. They use a new selection strategy based on benefit and overhead in load balancing algorithm. The measurement results of the experiment show that it can improve mean response time of jobs and resource utilization of systems substantially.

Abstract:Histogram is the commonest method for the size estimation of query result in many commercial database systems. Several histogram methods presented in the past have certain limitations in practicability due to their lack of guarantee of the accuracy of the estimation. In this paper, the author presents two kinds of new histogram methods, which are easy to use and can guarantee the accuracy of the estimation. Otherwise, the effect of different data distributions to histograms is discussed in the paper, and some important parameters of data distribution are introduced to help produce better histograms.

Abstract:As an extension of VoD (video-on-demand) system, DMIoDS (distributed multimedia information_on_demand system) has become a new application field of multimedia information processing technology. By analyzing several key problems involved in designing the DMIoDS, the authors present a kind of abstract hierarchical functional model of DMIoDS in this paper. According to the functional model, they present a prototype system and implement it.

Abstract:A model for computer network protocol specification and automatic generation is described in this paper. An extended attribute grammar to specify a protocol is introduced, on which the representations about parallelism, synchronization and timing in protocols are emphasized. Then, a protocol development environment is illustrated, and some referring issues, especially a semantics evaluation algorithm, are investigated.

Abstract:Mining association rules is an important data mining problem. There have been many algorithms proposed for efficient discovery of association rules in large databases. However, very little work has been done on maintenance of discovered association rules. When users interactively mine association rules, they may have to continually tune two thresholds, minimum support and minimum confidence, which describe the users' special interestingness. In this paper, two incremental updating algorithms——IUA and PIUA are presented for such efficient maintenance problem of association rules.

Abstract:In this paper, the authors established a mathematical structure of asymptotically dominating equivalence classes of computational complexity functions and studied the properties of the structure systematically. Applying the methods and techniques of recursion theory to the study of this structure, they proved that the structure has strongly Friedberg-Muchnic property and is dense under the sense of partial ordering.

Abstract:In this paper, the authors discuss some drawbacks of several kinds of existing logic to handle inconsistency in knowledge bases at first, and then give the definition of over-determined structure, over-determined model and over-determined semantics. They also describe in detail the computation procedure for over-determined-semantics. As a result, a kind of logic for handling inconsistency is presented. Finally, a discussion on the over-determined logic is concluded with.

Abstract:In this paper, a new integrated scheme for scheduling real-time traffic and cell loss control based on the queue length threshold is proposed and modeled using stochastic Petri nets. The connection-oriented nature of ATM networks can be kept and no resequencing device is needed in this new scheme. This large model of the scheme is decomposed into near-independent submodels to evaluate the model performance.