Advanced Materials Research Vols. 532-533

Abstract: Considering different quantum bit having different effective intensity in chromosome evolution, a novel quantum genetic algorithm based on potential is proposed. It makes the magnitude of rotation angle depending on the potential of a quantum bit. It generates the orientation of rotation angle according to the total potential of quantum bit in the chromosome. The character of quantum entangled interference based on potential is introduced. And convergence analysis and rationality analysis are implemented. Experimental test shows that, it can obtain better convergence rate and have less runtime on smaller population size and shorter chromosome length.

Abstract: Image denosing is the first preprocessing step in dealing with image processing where the overall system quality should be improved. So it is a key issue in all image processing researches. Over the past years, fractal-wavelet transforms were introduced in an effort to reduce the blockiness and computational complexity that are inherent in fractal image compression. The essence of fractal image denosing is to predict fractal code of a noiseless image from its noisy observation. From the predicted fractal code, we can generate an estimate of the original image. In the paper, we show how well fractal-wavelet denosing predicts parent wavelet subtrees of the noiseless image. The performance of various fractal-wavelet denosing schemes is compared to that of some standard wavelet thresholding methods. From the several of experimental results, these fractal-based image denosing methods are quite competitive with standard wavelet thresholding methods for image denosing.

Abstract: EM algorithm is a common method to solve mixed model parameters in statistical classification of remote sensing image. The EM algorithm based on fuzzification is presented in this paper to use a fuzzy set to represent each training sample. Via the weighted degree of membership, different samples will be of different effect during iteration to decrease the impact of noise on parameter learning and to increase the convergence rate of algorithm. The function and accuracy of classification of image data can be completed preferably.

Abstract: In this paper, we propose a new method for shape modification of NURBS curves. For a given NURBS curve, we modify its one or more weights so that the curve passes through the point specified in advance. We convert this into an optimization problem and solve it by genetic algorithm. The experimental results show the feasibility and validity of our method.

Abstract: K Nearest Neighbor (kNN) is a commonly-used text categorization algorithm. Previous studies mainly focused on improvements of the algorithm by modifying feature selection and k value selection. This research investigates the possibility to use Jensen-Shannon Divergence as similarity measure in the kNN classifier, and compares the performance, in terms of classification accuracy. The experiment denotes that the kNN algorithm based on Jensen-Shannon Divergence outperforms that based on Cosine value, while the performance is also largely dependent on number of categories and number of documents in a category.

Abstract: Computing with words is a methodology in which concepts, linguistic values or words, instead of numbers for computing and reasoning. Cloud model is a novel cognitive model for uncertain transformation between linguistic concepts and quantitative values. It should go without saying that computing with words based on cloud model is an important research study direction. After the algebra operations rule of cloud model is introduced, this paper presents a simple statistical algorithm to solve the computational problem of algebra operations for cloud model. Then, it proves that the statistical algorithm could reach any given precision level as long as cloud drops increase. Finally, we analyze the character of the statistical algorithm and the relationship with the operations rule for cloud model.

Abstract: we have used the metaphor of ant colonies to define "the Ant system", a class of distributed algorithms for combinatorial optimization. In this paper we analyze some properties of Ant-cycle, the up to now best performing of the ant algorithms we have tested. We report many results regarding its performance when varying the values of control parameters and we compare it with some FEM algorithms. And in accordance with treatment principles, the microstructure of the alloy is simulated.

Abstract: With the probability rules and the degree of coverage of elements in the partition set, and the combination of the Fuzzy Set Theory and Rough Set Theory, a new extension of Fuzzy Rough Set theory was proposed. It is defined the maximum of Rough Set membership function, the minimum of Rough Set membership function, the average of Rough Set membership function, the upper minimum of Rough Set membership function and the lower maximum of Rough Set membership function. The properties of the extension for the Rough Set membership functions are also given. This made Fuzzy Set Theory and Rough Set Theory complemented each other and provided a new way to deal with incomplete data.

Abstract: In this paper, α - limited dominance relation is defined for the classification of universe in the generalized incomplete information systems in which lost and “do not care” unknown attribute values coexist. Based on the α - limited dominance relation, an extended rough set model is formulated. Moreover, concepts of lower and upper approximations are studied as well as their properties. Finally, an example of evaluation of teaching quality is used to demonstrate that the proposed extended rough set model is feasible and effective in dealing with imprecise and uncertain knowledge.

Abstract: In signal processing, no matter the classical Fourier transform or the new wavelet transform is essentially related to kernel theory. The paper gives a reproducing kernel and its form is very simple. And by this kernel, the spline interpolation method can be constructed. Meanwhile, its best approximation property is proved. The numerical experiment results show that the interpolation method is convenient for numerical calculation and it has features of less calculation, higher approximation precision. Specially, the method has more advantages relative to Fourier analysis. That is, needless to prefilter and boundary extension, it can resolve ‘jumps’ at the edges in signal processing. And a new idea for signal simulation is put forward and the reproducing kernel theory is further enriched.