A Self-Organizing RBF Neural Network Based on Distance Concentration Immune Algorithm

Radial basis function neural network(RBFNN) is an effective algorithm in nonlinear system identification. How to properly adjust the structure and parameters of RBFNN is quite challenging. To solve this problem, a distance concentration immune algorithm(DCIA) is proposed to self-organize the structure and parameters of the RBFNN in this paper. First, the distance concentration algorithm, which increases the diversity of antibodies, is used to find the global optimal solution. Secondly,the information processing strength(IPS) algorithm is used to avoid the instability that is caused by the hidden layer with neurons split or deleted randomly. However, to improve the forecasting accuracy and reduce the computation time, a sample with the most frequent occurrence of maximum error is proposed to regulate the parameters of the new neuron. In addition, the convergence proof of a self-organizing RBF neural network based on distance concentration immune algorithm(DCIA-SORBFNN) is applied to guarantee the feasibility of algorithm. Finally, several nonlinear functions are used to validate the effectiveness of the algorithm. Experimental results show that the proposed DCIASORBFNN has achieved better nonlinear approximation ability than that of the art relevant competitors.

Study on Phylogenetic Relationships of Five Breeds of Pigs by Random Amplified Polymorphic DNA(RAPD)

RAPD was used to study the genetic divergency and phylogenetic relationships of five breeds of domestic pigs,including Min pig,Duroc,Yorkshired,Landrace and Junmu I pig.We selected fourteen primers from eighty random primers,caculated genetic distance index matrix and constructed phylogenetic tree with UPGMA methods.Genetic distance index matrix indicated that the genetic relationship between Junmu I pig and Landrace was the closest and the farthest between Duroc and min pig.

THE APPLICATION OF PATTERN RECOGNITION TECHNIQUES IN FAULT DIAGNOSIS OF MACHINERY EQUIPMENT

In this paper, the characteristics of vibration signal of machinery in different running conditions are statistically analysed, and some moments of statistical distribution of signals are selected as the eigenvector to condense the state information. Here, we divide the states of machinery into two: 'good' and 'faulty', and the pattern recognition techniques are used to classify the running conditions of machinery. At the end of this paper, the authors present some test data, and from the results obtained, it's verified that the eigenvector selected is reliable and sensible to faults. And the results also show the effectiveness of classification rule.

Capillary bridges and capillary forces between two axisymmetric power-law particles

Capillary interactions are fundamentally important in many scientific and industrial fields. However, most existing models of the capillary bridges and capillary forces between two solids with a mediated liquid, are based on extremely simple geometrical configurations, such as sphere-plate, sphere-sphere, and plate-plate. The capillary bridge and capillary force between two axisymmetric power-law profile particles with a mediated constant-volume liquid are investigated in this study. A dimensionless method is adopted to calculate the capillary bridge shape between two power-law profile particles based on the Young-Laplace equation. The critical rupture criterion of the liquid bridge is shown in four forms that produce consistent results. It was found that the dimensionless rupture distance changes little when the shape index is larger than 2. The results show that the power-law index has a significant influence on the capillary force between two power-law particles. This is directly attributed to the different shape profiles of power-law particles with different indices. Effects of various other parameters such as ratio of the particle equivalent radii, liquid contact angle, liquid volume, and interparticle distance on the capillary force between two power-law particles are also examined.