Abstract:The Unified Modeling Language (UML) is a general purpose visual modeling language that is designed to specify, visualize, construct and document the artifacts of software systems. UML sequence diagrams describe a collaboration of interacting objects, where the interactions are exchanges among communicating entities in real time and distributed systems. Like any other aspect of the specification and design process, the specifications in UML sequence diagrams are susceptible to errors, and their analysis is necessary.A tool for timing analysis of UML sequence diagrams is described in this paper.

Abstract:Computation partitionings (CP) for the data parallel statements in a program have a dramatic impact on its performance. Although the problem has been widely studied, previous approaches focus on improving spatial locality of the chosen CP. A time slicing optimization framework is presented, which integrates many important optimization strategies, to select optimal CPs for parallel loop constructs. In the framework, a CP is represented by a directed graph, which not only represents a mapping of the operations in aparallel state-ment into processprs,but also specifies the dependency constraints for operations in different processors.This approach is to evaluate the efficiency of each CP choice and to find the one with the best overall execution time.The evaluation method synthesizes the four aspects of load-balance,operation-independence between processors, spatial locality and temporal locality for each CP.The framework has been implemented in a HPF compiler p-HPF for FORALL construct.Experimental results show that the framework is of generality with desired speedups for a wide variety of data-parallel applications.With a very little modification,it can also be applied to many other kinds of data-parallel statement.

Abstract:Without software metrics, software would be error prone, expensive and with low quality. Cohesion is one of the most important factors for software quality as well as maintainability, reliability and reusability. The module of poor quality should be a serious obstacle to system quality. In order to design and maintain good quality software, software managers and engineers inevitably need to introduce cohesion metrics to measure and produce desirable software. In this paper, we propose a function oriented cohesion metrics is proposed.A series of experiments are shown to support the cohesion metrics,and aset of properties to evaluate the proposed cohesion metrics.Therefore,a well-defined, well-experimented and well-evaluated cohesion metrics is proposed to indicate cohesion strength and thus improve software quality.Furthermore,this cohesion metrics can be easily incorporated into software CASE tool to help software engineers to ensure software quality.

Abstract:Based on the backward mapping technique, an efficient image-based rendering algorithm for novel view generation from multiple reference images in a static scene is presented in this paper. It successfully fills the holes in the derived image both occured when object surfaces are magnified in the novel view and when occluded surfaces become visible. This algorithm is more efficient by the heuristic technique proposed in this paper which gets displacement values used during backward mapping process from single main reference image under the observation that, the shape of space surfaces often changes smoothly. Comparing with the usual forward mapping methods, this algorithm generates derived images with less errors and tackles the problems associated with multiple reference images.

Abstract:Differentiated service (DiffServ) architecture has become one of the research hotspots on computer networking in the last two years. Its intention is to provide quality of service for users at coarse granularity level. In contrast to integrated service architecture, DiffServ is not only more scalable, but also easier to be deployed in traditional packet switching networks. In this paper, the packet-marking algorithm, one of the key mechanisms in DiffServ is studied. A new packet-marking algorithm called FMPA (fair marking packet algorithm) is presented. The new algorithm is compared with the existing proportional pachet-marking algorithm by simulation. In addition, a packet remarking algorithm is proposed. By using it, the packet's original semantic can be kept to the largest degree. The simulation result shows it in this paper.

Abstract:In recent years, relevance feedback technique has become an active research method in image retrieval . A self-learning algorithm of image retrieval using forward propagation neural network is proposed in this paper. During the interactive retrieval process, users can mark positive images similar to the query image. Then the algorithm constructs a forward neural network and retrieves again based on the learned neural network. The experimental result over 9 918 images shows that the proposed approach greaty reduces the user's effort of composing a query and representing a concept. During the interactive learning and retrieval process, more and more correct images can be found in the anterior result. This approch is robust to various kinds of feature representation and simiarity distance formulas.

Abstract:Maintaining authenticity for multicast communication is inherently more complex than for unicast, which dues to the dynamic group and unreliable delivery. In the implementation of multicast authentication, the major obstacles lie on the lower signature rate and larger signature size overhead. To solve this problem, by using a new authentication technique called authentication matrix, a new signature scheme is proposed for large and dynamic multicast groups that can be used on the unreliable delivery network. Comparing with the existing schemes, in this scheme, the signature size overhead is much smaller and the signature rate is much higher, and it can provide the non-repudiation serve. It should have application in many practical fields such as multimedia data transmission, live multi-party conferencing and long-distance education.

Abstract:Polygon simplification in scaleless GIS (geographic information system) is based on the line generalization of polygon arcs, but there may be topological errors in output data resulting from the direct line generalization of polygon's arcs, and the major representation of the error is the intersection of the simplified polygon s arcs. The previous research did not attach importance to this problem. A method of polygon simplification based on Strip-Tree in scaleless GIS is presented in this paper through an all-sided analysis of this problem. Maintaining the correctness of the polygon layer's topoloty, it simplifies the polygon layer according to the output scale. This method has been implemented into an 863 project, and the experimental results show that this method can simplify the polygon layer with high efficiency and can maintain the shape of polygon very well.

Abstract:In web mining, applying association rule discovery can discover the association between different web pages accessed by users. Because there is the rich structure information in the website and the access of the users conforms to some kinds of sequences, a new approach is presented in this paper to discover the association between the access sequences, which is the sequence association rule discovery. In this approach, first the Log is mined in the web server to get the user access transactions, and then according to the regular grammar, a new user access transaction grammar is defined in order to get the sequence access transactions from the user access transactions. Subsequently, the association rule discovery algorithm is employed to discover the sequence association rules. To evaluate these rules, the mutual information is proposed. The result of this approach can help the designer of the website to understand the user access patterns better, and according to this result, the designer can adjust the structure of the web site.

Abstract:On the base of analysis classic methods, the morphologic addition algorithm for convex polyhedron is predigested to the morphologic addition of faces in polyhedron. The concept of reference plane is introduced with the model of normal vector sphere for 3D objects, the convex polygons in 3D space are divided into two parts by the reference plane and morphologic addition of each corresponding part is calculated. Then the repetitionary faces and edges are deleted, a fast morphologic algorithm for convex polyheron is presented. The experimental results show that this method is 6～10 times faster than classic methods and works well.

Abstract:The current researches of knowledge discovery are introduced under the environment of incomplete data, and then an approach is presented for knowledge discovery in incomplete databases through the two parts. Firstly, the method of how to guess the missing data is in detail discussed and the definition as well as the mining method of distance based association rule is given. Then based on this, an algorithm is described in detail for discovering knowledge in incomplete databases, the complexity of the algorithm is analyzed and the experimental results are also given. The paper finally concludes with a comparison of the approach proposed in the paper with other related ones.

Abstract:The general packet radio service (GPRS) aims at providing predictive QoS to user by using QoS management mechanisms similar to those employed by ATM. It shows that the measurement based admission control algorithm has better adaptability and higher efficiency than other call admission control algorithms in many literatures. In this paper, based on the enough analysis on characteristic of GPRS, a measurement based admission control algorithm is presented fitting for GPRS using a model for the worst-case delay of priority queue, equivalent approximate leaky bucket and delay normalized. By simulation, the proposed algorithm can satisfy the required GPRS QoS while achieving high sources utilization rates of about 80%.

Abstract:Data fusion has been widely applied in the remote sensing research field. Principal component analysis (PCA) is one of the standard methods for data fusion. In this paper, a new algorithm--adding wavelet coefficients principal component analysis (AWPCA) is presented, which is based on principal component analysis (PCA) and is gotten from combining PCA and wavelet transform. The experimental results demonstrate that the higher quality image is obtained by AWPCA than by IHS and PCA mergers. AWPCA can be also applied in other fields where the high-resolution image is required.

Abstract:A discrete-HMM training algorithm which has strong ability of pattern classification is presented in this paper. By VQ (vector quantization) technique, this algorithm trains data from all speakers in mixed mode to generate the speaker characteristic pattern, which includes features of all speakers. By substituting the VQ code\|book in conventional discrete-HMM with characteristic pattern, the ability of pattern classification for observation symbol sequence is enhanced, therefore the classifying ability of discrete-HMM is improved. The experimental results show that the algorithm can improve the system's recognition performance, and reduce the dependence extent of HMM on the scale of training set. Moreover, the calculation quantum of this algorithm in recognition stage is obviously less than that of conventional HMM training algorithm, therefore it has higher practical value.

Abstract:A generalized morpho-interpolation method for nonrigid body motion is proposed based on morpho translation. The problem of morpho-interpolation between two nonhomotopy polygons (including concave polygon and holey polygon) is solved by decomposing polygon into several individual convex subpolygons and constructing the matching of two convex subpolygons. It is proved that nonrigid body motion can be divided in to the composing of nonrigid body metamorphosis and rigid body rotating. The principle of the morpho-interpolation is also proved and the effects of morpho-interpolation with different convex decomposing are discussed. Experimental results show that this method can get the natural, higher quality metamorphosis sequence with simple computation and can be used in description of nonrigid body motion and automatically interpolation between two deyframes in 2D animation.

Abstract:Based on maximum rank distance codes, a new kind of Stern scheme is proposed in this paper, security of this scheme is discussed and it is proved that this scheme is secure when the parameters are selected properly .

Abstract:Priority is an important approach to the control of concurrent systems. It is often applied to solve the conflict problems in the design of concurrent systems. In this paper, the concept of dynamic priority systems is developed based on bounded P/T systems and the interleaving and the true-concurrency semantics are provided for them respectively. Dynamic priority systems can be used as both modeling tools to build concurrent and distributed systems and semantic foundations to define prioritised operators in programming languages. Finally, the Occam language with a dynamic prioritised operator based on the concept of dynamic priority systems is extended.

Abstract:The scalability of system is a puzzle to design high performance computers. NUMA (non_uniform memory architecture) architecture is proposed for scalable system underlying a shared memory pattern. Research and practice results show that the scalability of system tightly correlates with that of the operating system. Usually, two structures of operating system are employed in multiprocessor systems, monolithic based on shared memory and multi kernel based on message passing. However, both of them cannot fully fit scalable parallel computers, especially MPP systems with NUMA architecture. In this paper, the problems in two structures are analyzed, a new structure, multi-virtual-space and multi-mapping with active message, is proposed. The test and practice results show that the new structure is successful in the scalability of system. This paper introduces detailedly the implementation of the structure in a MPP system with NUMA architecture.

Abstract:Using complex analysis and curve integration, the construction of PH quintic which satisfies Hermite interpolation conditions is studied in this paper and its corresponding Bézier representation is derived. The PH quintic has continuous unit tangents and signed curvature, and its arclength function is the polynomial of its parameter. The PH quintic has offset curve that admits exact rational algebraic representation, intuitive geometrical interpretation and can flexibly deal with inflection point.

Abstract:In this paper, the semantics of Verilog program are studied in a discrete continuous hybrid time model, a hybrid interval is denoted as the description of a run of Verilog program. And an extended ITL is presented to describe this kind of hybrid interval. This semantics not only considers the ultimate execution effects of all kinds ingredients of language, but also gives their temporal characrestics.