Database designers and students of computer science in tertiary institutions find it difficult to learn database theory in which database decomposition is a major aspect. This work presents a visualization tool to simplify the study of decomposition algorithms behaviours which would assist the students and database designers to decompose any relational schema with its functional dependencies in case it violates third normal form or Boyce Codd Normal Form. Students perceptions are captured via translating decomposition algorithms into visualization tool for entering relational attributes with their functional dependencies. Students are requested to move an arrow from one normal form to another and the system specifies a green light as an indication of normalized form or red light as an indication of violation of normal form for learning purposes. Students use the tool to understand the abstract knowledge of decomposition techniques and easily use it to decompose any relation if it violates either third normal form or Boyce Codd normal form. The effectiveness of the tool has been evaluated in surveys and the students generally viewed it more positively than conventional classroom teaching.

In this paper, two kinds of synchronization between two complex networks with non-delayed and delayed coupling are discussed by the pinning control method, that is, inner synchronization and outer synchronization. Based on the Lyapunov stability theory and linear matrix inequality (LMI), some sufficient conditions for the synchronization are derived by adding linear feedback controllers to a part of nodes, and the linear feedback controllers are designed. Under suitable conditions, not only inner synchronization but also outer synchronization can be asymptotically achieved. Numerical simulations are presented to show the effectiveness of the proposed synchronization scheme.

Load balancing algorithms goal is to keep all nodes normally loaded through migration of modules from heavy loaded nodes to lightly loaded nodes. In addition, load balancing must involve low communication overhead and respond quickly to load imbalance in the system. In previous load balancing algorithms, classification of node as heavy, light or normal loaded node is done by using concept of threshold level, which is fixed and predefined. Now, today scenario is to change the status of nodes dynamically according to the state of system. So that, in this paper we proposed an algorithm for load balancing using fuzzy clustering; which improves the performance of the system without pre-defining the threshold values. The proposed algorithm is compared with other existing algorithms and is found to be fast and efficient in reducing load imbalance in Peer to Peer system.

Document classification and document summarization have a fairly indirect relation as document classification fall into classification problems as opposed to document summarization, where it is treated as a problem of semantics. A major part of the summarization process is the identification of the topic or topics that are discussed in a random document. With that in mind, we try to discover whether document classification can assist in supervised document summarization. Our approach considers a set of classes, in which a document may be classified in, and a novel summarization scheme adapted to extract summaries according the results of the classification. The system is evaluated against a number of supervises and unsupervised approaches and yields significant results.

As the use of digital imaging is on the rise, compression of acquired digital image data is becoming more and more important to cope with the storage requirements. One of the challenges of compression is to compress images with high efficiently while preserving critical data from getting permanently lost in reconstructed images. In this research we introduce a proposed algorithm for image compression, based on the Minimize-Matrix-Size Algorithm for coding and Limited Sequential Search- Algorithm (LSS-Algorithm) for decoding. The proposed algorithm starts by using single stage Discrete Wavelet Transform, to decompose an image into four subbands; low frequency and high frequencies. Each "n × n" of the low-frequency subband (LL) are transformed by using two dimensional DCT, store all DC coefficients in different matrix called DC-Matrix, and the remain AC coefficients are stored in different matrix called AC-Matrix, then applying Minimize-Matrix-size algorithm for the AC-Matrix, to convert each group of AC coefficients into single floating point value. The DC-Matrix transformed again by DWT, and apply Minimize-Matrix-Size algorithm on it. LSS-Algorithm which is represents decoding; DC-Matrix and AC-Matrix, this algorithm is used to estimate original values by using iterative method, which is depends on the probability of the data of the AC-Matrix and DC-Matrix. Our compression algorithm proved good compression ratio, and compared with JPEG and JPEG2000 depending on the PSNR and HVS.

To keep the key idea of rough set and the representation of information in rough set theory, empty representation is processed properly and three new forms of rough approximation sets are defined as a generation of general binary relation based rough set model. Moreover, the properties of approximation operators in these new rough sets are discussed. The relations among them are studied in this paper. In addition, examples are arranged to interpret what we studied in this paper.

In this paper, a new nonlinear conjugate gradient methods with wolfe type line search for solving unstrained optimization problems is proposed. The direction generated by the new methods produce sufficient descent search direction. Under some conditions, we give the global convergence results for the new nonlinear conjugate gradient methods with Wolfe type line search. Finally, the numerical results show that the new method is also efficient for general unconstrained optimizations.

A new kind of quasi-quartic trigonometric polynomial base functions with a shape parameter λ over the space Ω=span {1, sint, cost, sint2t, cos2t} is presented, and the corresponding quasi-quartic trigonometric Bézier curves and surfaces are defined by the introduced base functions. The quasi-quartic trigonometric Bézier curves inherit most of properties similar to those of quartic Bézier curves, and can be adjusted easily by using the shape parameter λ. The jointing conditions of two pieces of curves with G2 and C4 continuity are discussed. With the shape parameter chosen properly, the defined curves can express exactly any plane curves or space curves defined by parametric equation based on{1, sint, cost, sint2t, cos2t} and circular helix with high degree of accuracy without using rational form. The corresponding tensor product surfaces can also represent precisely some quadratic surfaces, such as sphere, paraboloid, cylindrical surfaces, and some complex surfaces. The relationship between quasi-quartic trigonometric Bézier curves and quartic Bézier curves is also discussed, the larger is parameter λ, and the more approach is the quasi- quartic trigonometric Bézier curve to the control polygon. Examples are given to illustrate that the curves and surfaces can be used as an efficient new model for geometric design in the fields of CAGD.