Applied Mechanics and Materials Vols. 687-691

Abstract: Spectral clustering algorithm is a kind of clustering algorithm based on spectral graph theory. As spectral clustering has deep theoretical foundation as well as the advantage in dealing with non-convex distribution, it has received much attention in machine learning and data mining areas. The algorithm is easy to implement, and outperforms traditional clustering algorithms such as K-means algorithm. This paper aims to give some intuitions on spectral clustering. We describe different graph partition criteria, the definition of spectral clustering, and clustering steps, etc. Finally, in order to solve the disadvantage of spectral clustering, some improvements are introduced briefly.

Abstract: In this paper, steady-state anisotropic heat conduction equation can be converted into the first kind integral equation, then modified quadrature formula based on trapezoidal rule is used to deal the integrals with singular kernels. In addition, Sidi transformation is applied to remove the singularities at concave points in concave polygons. This technique improves the accuracy of numerical solutions of the heat conduction equation. Numerical results show the convergence rate of the proposed method is the order three.

Abstract: Fuzzy neural networks is a powerful computational model, which integrates fuzzy systems with neural networks, and fuzzy perceptron is a kind of this neural networks. In this paper, a learning algorithm is proposed for a fuzzy perceptron with max-product composition, and the topological structure of this fuzzy perceptron is the same as conventional linear perceptrons. The inner operations involved in the working process of this fuzzy perceptron are based on the max-product logical operations rather than conventional multiplication and summation etc. To illustrate the finite convergence of proposed algorithm, some numerical experiments are provided.

Abstract: In this paper, discusses the school bus problem, given the mathematical model of route optimization, put forward an improved cuckoo search algorithm. Finally, take school bus routes running problem of Gansu Normal College for Nationalities as an example, the algorithm was experimented and the experimental results show that the algorithm to be successful.

Abstract: In classical harmony search algorithm, only one harmony vector is obtained in each of iteration, which affects its search ability. We propose an improve harmony search algorithm in this paper. In our approach, the number of harmony vectors obtained in each of iteration is equivalent to the population size, and all newly generated harmony vectors are put into the harmony memory array. Then, all harmony vectors are sorted by descending order of the fitness, and the first half individuals are served as the next generation of populations. Experimental results show that our approach is obviously superior to the classical one under the same iteration steps and the same running time, which reveals that our approach can effectively generate the excellent individuals approximating the global optimal solution and enhance the optimization ability of classical harmony search algorithm.

Abstract: To address the fitting spherical surface and evaluating sphericity error, a mathematical model based on the minimum zone principle is presented. And the presented model is answered by GA. An example shows the performance of the proposed method by comparison with the methods based on the least square principle.

Abstract: Rough set theory and grey theory have the same advantage of processing inaccuracy data, so one fusion algorithm based on them is proposed. The attribute reduction algorithm of rough set theory can reduce the decision table of fault diagnosis, and grey theory can predict the fault based on the new reduced decision table. Then it is verified in some aero radio equipment, and the results indicate that the accuracy of fault prediction is quite higher, which provides the foundation to improve the equipment reliability and maintainability.

Abstract: In this paper, we propose a fast multiobjective particle swarm optimization algorithm (called CBR-fMOPSO for short). In the algorithm, a case-based reasoning (CBR) technique is used to retrieve history optimization results and experts’ experience and add them into the population of multiobjective particle swarm optimization algorithm (MOPSO) in dynamic environment. The optimal solutions found by CBR-fMOPSO are used to mend the case library to improve the accuracy of solving based on CBR in next solving. The results from a suit of experiments in electric furnaces show that the proposed algorithm maintains good performances however the environment changes.

Abstract: Data mining and optimize traditional Chinese medicine (TCM) prescription compatibility based on wavelet denoise spectral and partial least squares (WDS-PLS). Method: First of all, experimental design: with reference to the original formula, the herbal medicines in a prescription designed nine formula based on mixing uniform design; Secondly, obtain experimental data and data standardization; Finally, mathematical modeling, data mining and optimize TCM prescription compatibility base on WDS-PLS.Results: gain the regression coefficient and equation, VIP sorting, loadings Bi plot, and seek out the optimized direction of the prescription. Conclusion: the method data mining and optimize the compatibility of the dachengqi decoction is feasible and effective.

Abstract: In order to test the Simulink models, this paper focused on the model-based testing for Simulink based on mutation testing. Considering the situation that the present model-based test adequacy criteria are imperfect, we propose a mutation testing process for Simulink models. Based on studying the application of mutation testing technique to Simulink models, some improvements on mutation operators for Simulink models are presented. The experimentation results verified the effectiveness and correctness.