Abstract:In many models of optical routing, a set of communication paths (req uests) in a network are given, and a wavelength must be assigned to each path so that paths sharing an edge receive different wavelengths. The goal is to assign as few wavelengths as possible, in order to make as efficient use as possible of the optical bandwidth. Much work in the area has considered the use of wavelen gth converters: if a node of a network contains a converter, any path passing th rough this node may change its wavelength. Having converters at some of the nodes can reduce the number of wavelengths down to congestion bound. Thus Wilfong and Winkler defined a set S of nodes to be sufficient if, placing converters at the nodes in S, every set of paths can be routed with a number of wavelengths equal to its congestion bound. In this paper, the minimum sufficient set problem in bi-directed networks is studied. The problem is transformed into minimum vertex cover problem and some algorithms are developed for the problem.

Abstract:In order to solve switching regression problems, many approaches have been investigated. In this paper, anintegrated fuzzy clustering algorithm GFC that combines gravity-based clustering algorithm GC with fuzzy clustering is presented. GC, as a new hard clustering algorithm presented here, is based on the well-known Newton's Gravity Law. The theoretic analysis shows that GFC can conve rge to a local minimum of the object function. Experimental results show that GFC for switching regression problems has better performance than standard fuzzy clustering algorithms, especially in terms of convergence speed.

Abstract:The relationship between the total template dependencies and the total join dependencies is probed into by means of abstract algebra. First, two equ ivalence relations are defined in the set of the total template dependencies and the set of the total join dependencies respectively. The equivalence relations regard the dependencies that function is the same as equivalent dependencies. Th en, it is proved that two quotient sets under two equivalence relations constitute monoids respectively and there is an isomorphism mapping between the monoids, which shows that the class of the total join dependencies is essentially identical with the class of the total template dependencies. Finally, an interesting result about the total acyclic join dependencies is given. The relevant results will play active role in designing relational database schemes.

Abstract:Performance issues are vital to the success of many interactive soft ware systems, including software development tools and Web-based tools. However, the issue of performance engineering is still not emphasized in the process of software development. The key aim of this paper is to illustrate an effective me thod to achieve satisfactory performance for interactive software after careful design. This paper presents how to apply the software performance engineering (SPE) method by focusing on performance estimation at the design stage, and itsef fect on determining implementation approaches, in the development of interactive software tools. In addition to the rigid quantitative estimation method origina ted from SPE, this paper argues that the performance can also be cost-effectively estimated either semi-quantitatively or non-quantitatively. With the experience results described in this paper, it is suggested that it may not be compulsory to achieve direct quantitative performance estimation from environment specific ations for every software development as advocated by the SPE method. This paper demonstrates a combination of analytical and experimental approaches to assessing the performance at early stages in development of software tools. It is hoped that this can reflect the essence of good experimental computer science.

Abstract:With the rapid growth of online text information and user accesses, query-processing efficiency has become the major bottleneck of information retri eval (IR) systems. This paper proposes two new full-text indexing models to impr ove query-processing efficiency of IR systems. By using directed graph to repres ent text string, the adjacency matrix of text string is introduced. Two approach es are proposed to implement the adjacency matrix of text string, which leads to two new full-text indexing models, I.e., adjacency matrix based inverted file and adjacency matrix based PAT array. Query algorithms for the new models are dev eloped and performance comparisons between the new models and the traditional models are carried out. Experiments over real-world text collections are conducted to validate the effectiveness and efficiency of the new models. The new models can improve query-processing efficiency considerably at the cost of much less amount of extra storage overhead compared to the size of original text database, so are suitable for applications of large-scale text databases.

Abstract:In this paper, hierarchical mass-assignment fuzzy systems of two types are presented, based on mass assignment theory. It is constructively provedt hat hierarchical mass-assignment fuzzy systems of these two types are also unive rsal approximators. Because of the fact that the number of rules in type 1 hiera rchical mass-assignment fuzzy systems increases linearly with the number of input variables and that fuzzy systems are added up to type 2 hierarchical mass-assignment fuzzy systems in terms of different accuracy requirements. These two types of systems can be effectively used to overcome rule-explosion problem, that is , the number of rules increases exponentially with the number of input variables.

Abstract:The crucial technologies related to personalization are introduced in this paper, which include the representation and modification of user profile, the representation of resource, the recommendation technology, and the architecture of personalization. By comparing with some existing prototype systems, the key technologies about how to implement personalization are discussed in detail. In addition, three representative personalization systems are analyzed. At last, some research directions for personalization are presented.

Abstract:Disconnected operation is a challenging issue in mobile computing. Data hoarding is the process of prefetching important data into local cache for future operation. The result of hoarding process will dramatically affect the performance of disconnected operation. A low-cost general hoarding algorithm is introduced to exploit relationship among the data items simultaneously when processing data requests and to select data for hoarding automatically. Simulation results show that this algorithm effectively improves cache hit rate in disconnected operation. And it is especially applicable for handheld mobile devices with low storage capacity and solw computing speed.

Abstract:Similarity search is a very important problem in data mining. It retrieves similar objects in database and finds proximity between objects. It can be applied to image/picture databases, spatial databases, and time-series databases. For Euclid space (a special metric space), similarity search algorithms based on R-tree are efficient in low-dimensional space, but degenerate into linear scan for high-dimensional space. This phenomenon is called dimensionality curse. This paper presents a new partition and index method of metric space, rgh-tree which distributes and partitions objects by using distance information if objects with rew fixed reference. It produces a balance tree with no data overlay. In addition, an algorithm based on rgh-tree, which is suitable for similarity search in metric space, is presented in this paper. The algorithm overcomes the shortcomings of the exiting algorithms, which has less I/O cost and times of computing distance, with average complexity o(n^0.58).

Abstract:In this paper, an object-oriented domain engineering method, Jade Bird domain engineering method is proposed. In domain engineering, several systems in a domain are analyzed, and their commonalities and variabilities are identified. Through development for reuse, domain-specific components and the architecture are produced. Domain engineering approach helps successfully for software reuse. Based on the Jade Bird object-oriented method, this method defines activities and artifacts in every domain engineering stage, and gives guidelines for each stage and the activies in it.

Abstract:By lack of suitable abstractions of the resource model, it is difficult to apply well-established software development methods to Web applications. An object view model WebView is outlined and the model is implemented in this paper as a Web application framework. WebView uses object view as a uniform concept to model Web entities. By using the model in designing, it can enhance reusability and productivity in the Web application development. Based on the available Web technologies, the Web application framework can map object views into the HTML format.

Abstract:Security is an important obstacle to the application of mobile agent technology. After analyzing the security problems and current status of research work, several methods are proposed to solve some of the problems, and are implemented in the mobile agent system Mogent. These methods can prevent or solve some of the security problems.

Abstract:The biggest problem in content-based image retrieval (CBIR) is a big gap between high-level semantic contents and low-level features. As an effective solution, relevance feedback has been put on many efforts for the past few years. In this paper, a new relevance feedback approach with progressive learning capability is proposed. It is based on a Bayesian classifier and treats positive and negative feedback examples with different strategies. It can utilize previous users' feedback information to improve its retrieval ability. The experimental results show that this algorithm achieves high accuracy and effectiveness on real-world image collections.

Abstract:The decomposition methods are main family to train SVM (support vector machine) for large-scale problem. In many pattern classification problems, most support vectors?Lagrangian multipliers are bound, and those multipliers change smoothly during training phases. Based on the facts, an efficient caching strategy is proposed to accelerate the decomposition methods in this paper. Platt抯 sequential minimization optimization (SMO) algorithm is improved by this caching strategy. The experimental results show that the modified algorithm can be 2~3 times faster than the classical SMO for large real-world data sets.

Abstract:In curve and surface modeling, Bernstein polynomial compositions are widely used for various geometric operations. So it is important to investigate them both in theory and practice. The problems are investigated by using polynomial interpolation and symbolic computation, and the proposed method is applied for curve and surface cases. Compared with two existing methods, the proposed method has the advantages on computational cost, coding efficiency, storage cost. However its numerical accuracy is lower than the method based on the generalized de Casteljau algorithm.

Abstract:In this paper, a structural operational semantic model is presented for a core subset of Verilog, and the subset has the main features of Verilog such as event-driven computation, shared-variable concurrency, time-delay, and so on. And all the Verilog processes are seen as open systems in this operational semantic model, so a model of observation is provided for open Verilog processes, and use observation equivalence based on bisimulation to identify the equivalence between programs. The observation equivalence can be proved to be a congruence for all Verilog operators, so it provides a sound base for deriving the algebraic laws for Verilog processes.

Abstract:The color histogram based image retrieval method is simple and efficient but losing the spatial distribution information of the color. In this paper, a color-space feature based image retrieval method is presented. The content of one image is looked as the aggregation of some main objects, which can be obtained by image segmentation. The total similarity between two images is computed according to these main objects?color, location and shape features and then retrieval there images. The experimental results show that the method is more accurate and efficient in retrieving the user interested images than color histogram method when there are obvious objects in the image.

Abstract:In this paper, a mechanism of implementing visualization with level of detail (LoD) at multi-scale is proposed on spatial database. Based on the spatial index V-Reactive tree, different LoDs can be represented, at the same time generalization at multi-scale can be implemented. Furthermore, the perspective query algorithm can decrease the complexity of implementing the mechanism. According to the characteristics of 3D visualization, the index tree is optimized for better image quality. The experimental results show that the mechanism is fit for 3D GIS applications with mass data.

Abstract:In order to restore degenerative images, which are go short of priori knowledge about original images, and explore new ways of x-ray tomographic image reconstruction, the experience of Spall and Cristion抯 simultaneous perturbation stochastic approximation (SPSA) method is drawn on, and this algorithm is extended to the high order and multivariate case, then a new gradient approximation algorithm with stochastic perturbation is presented. This algorithm does not need either a priori knowledge or a posteriori probability, and has convergence with excellent stability. Comparative experiments show that this algorithm converges to visually good images with excellent stability for restoration and reconstruction of images.

Abstract:A method to translate Chinese text into Chinese sign language is introduced, and how to present gesture by a 3D virtual human, to help the deaf communicate with the hearing. In this method, an initiate database including all sign words motions is firstly created based on a motion capture method, where two data gloves and three 6-DOF position tractors are used to record the motion of a real person when he or she demonstrates each sign word. Then all motion data can be tuned accurately and easily by modifying the gestures just on some control points. After that, a database that contains of all sign words has been built up. Given a sentence of text, the motion of each word can be found in the database, concatenated one by one, and presented smoothly by a virtual presenter who is represented by a VRML virtual human model, then the sentence is translated into Chinese sign language. To help the deaf study, watch TV, and browse Internet, a system that can be used to translate Chinese text into Chinese sign language is developed in the PC/Windows/VC environment, and all most of generating signs by this system are very realistic and smooth as confirmed by a lot of teachers in deaf schools. The database of this system contains motions of all 5 596 Chinese sign words collected by a handbook of Chinese sign language.