Abstract: Improved learning algorithm for branch and bound for semi-supervised support vector machines is proposed, according to the greater difference in the optimal solution in different semi-supervised support vector machines for the same data set caused by the local optimization. The lower bound of node in IBBS3VM algorithm is re-defined, which will be pseudo-dual function value as the lower bound of node to avoid the large amount of calculation of 0-1 quadratic programming, reducing the lower bound of each node calculate the time complexity; at the same time, in determining the branch nodes, only based on the credibility of the unlabeled samples without the need to repeatedly carry out the training of support vector machines to enhance the training speed of the algorithm. Simulation analysis shows that IBBS3VM presented in this paper has faster training speed than BBS3VM algorithms, higher precision and stronger robustness than the other semi-supervised support vector machines.
Abstract: This paper introduced a computer simulation model for composite materials which was reinforced by spherical particles. We introduced its algorithm and visualize the model with different particle volume fraction. In order to evaluate the uniformity of the particle distribution, we estimated Particle Center Density and standard deviation of minimal sphere distance.
Abstract: A stochastic Von Neumann model to describe companies’ input-output process on mobile services is provided. Through the theories from input-output economics, an extended singular stochastic Von Neumann model on mobile service is researched. The problem of stability of this kind of stochastic Von Neumann model on mobile services is researched. A new mathematic method is applied to study the singular systems without converting them into general systems. The parameter uncertainties are considered. A new stability criterion for the extended stochastic Von Neumann model is given to ensure the stability of input-output model.
Abstract: According to the features of HVS, the algorithm uses region-split method to segment the remote sensing image into a large number of small regions. By integrating gray feature and spatial location of each region, NC is used to segment the image among regions from global view, by which the final segmented image can be generated. Experimental results show that comparing with the traditional NC, operating speed is significantly improved as getting close segmentation quality, and this is a kind of effective method of image segmentation.
Abstract: Data Mining system has a large number of data processing tasks. This paper mainly studies on task scheduling problem which takes the minimum total completion time based on the m equivalent processors and n tasks with priority. Firstly, we propose a data mining system multi-task scheduling model. Secondly, we adopt genetic algorithm solving this model. Due to initial seed may not be satisfied with the precedence constraint, this paper uses the inverse adjacency list to modify. Finally, we show this model and its algorithm are feasible and valid by the simulation experiment.
Abstract: Based on virtual prototyping technology, all mould vibration devices were framed in three-dimensional modeling, defining each components’ quality, center of mass, pairs connection between the components in the system, which representated truely mold four-link vibration device dynamic characteristic, implementing the dynamic behavior simulation. Kinematics law of the oscillation mechanism arose: Velocity and acceleration of the components in this vibration device were sine or cosine law functions.The angular velocity and angular acceleration of the components is proportional to amplitude. The angular velocity of the components is proportional to vibration frequency. The angular acceleration is proportional to the square of the changes in vibration frequency.
Abstract: Since conventional electro-mechanical converter of 2D digital valve had limitations of large inertia and low dynamic response, a novel low inertia rotary electromagnet with coreless rotor structure was proposed. The axial magnetic flux generated by permanent magnet and the radial magnetic flux generated by excitation coils were superimposed and then electromagnetic torque was produced to drive rotor to track the input signal simultaneously. Based on the approach of 3D finite element electromagnetic field simulation, the influence of key structure and operating parameters such as excitation ampere turns, working air gap, tooth height, tooth width and wall thickness of coreless rotor on torque-angle characteristic was discussed. The result obtained provides a certain reference value on the structural optimization of electro-mechanical converter of 2D digital valve.
Abstract: The present study was conducted to investigate the effect of seed priming techniques on germination characteristics of C. maxima Duch. cultivar (Beiguan). Treatments were combinations of 3 levels of priming (distilled water, NaCl and PEG6000) and non-priming (control) with 3 replications. Concentrations of NaCl solution were 50, 100, 150, 200 mmol•L-1, and polyethylene glycol (PEG) 6000 were 10%, 20%, 30%, 40%, respectively. Seeds were primed using the above priming materials for 24 hours at 20°C in the dark, respectively. The results showed that different priming techniques could have various effects on germination of Beiguan seeds. Hydropriming (distilled water), NaCl priming and PEG6000 priming (10%) all improved the germination characteristics of Beiguan, compared to the control. NaCl priming was more effective than hydropriming and PEG6000 priming and was the most successful technique in this study. These findings indicated that seed priming techniques could accelerate germination process and were simple and cheap, we should propose these methods to farmers.
Abstract: To overcome the shortcomings of the traditional passive-radar-seeker (PRS) for anti-decoy, a complex angle measuring method is proposed in this letter. The complex angle measuring method consists of monopulse angle and spatial spectrum estimation, two angle-measuring units. PRS can get the angle high-resolution features through the complex angle measuring method. To improve the precision and the stabilization of the angle-measure in the active-decoying environment, a grey processing approach is proposed in this paper. The methods of removing gross error and parameter estimation based on the grey entropy definition are proposed in the paper. Using grey data processing approach based on the definitions of grey distance measure and grey relation entropy, eliminate singularity and retain random error. Finally, the simulation results verify the feasibility and anti-decoy capacity of the complex angle-measuring method, and the grey processing approach can improve the precision and stabilization of angle-measurer in active-decoying environment, and can improve the precision of anti-radiation missiles (ARM).