A new learning method using prior information of neural networks

In this paper, we present a new learning method using prior information for three-layer neural networks. Usually when neural networks are used for identification of systems, all of their weights are trained independently, without considering interrelated weights values. Thus, the training results are usually not good. The reason for this in that each parameter has its influence on others during learning. To overcome this problem, we first give an exact mathematical equation that describes the relation between weight values given a set of data conveying prior information. The we present a new learning method that trains part of the weights and calculates the others using these exact mathematical equations. This method often a priori keeps the given mathematical structure exactly the same during learning; in other words, training is done so that the network follows a predetermined trajectory. Numerical computer simulation results are provided to support this approach. This work was presented, in part, at the Fourth International Symposium on Artificial Life and Robotics, Oita, Japan, January 19–22, 1999.     

H∞ ALMOST DISTURBANCE DECOUPLING FOR MIMO BILINEAR SYSTEMS

The H_∞ almost disturbance decoupling problem is considered. In this paper, a nonlinear design is proposed to find a state feedback controller for bilinear systems. The closed‐loop system is internally stable and achieves disturbance attenuation in nonlinear H_∞ sense. We defined a special form of Lyapunov function, which is constructed in terms of one or a set of positive definite constant matrices. If, except of the origin of system, the corresponding polynomial of the positive definite matrix (or several polynomials relevant to the positive definite constant matrices) has (have) no zero on a given subset of state space, then we can construct a controller to solve our problem. It is found that the controller structure could be complicated, but is feasible in computation and may require optimization technique to search the solution. We consider both SIMO and MIMO cases with illustrated examples.     

水和甲醇存在下MAMS的消失反应动力学研究

通过实验研究了甲基丙烯酰胺硫酸盐（ＭＡＭＳ）与甲醇、水进行反应的动力学，并对该反应的机理进行了探讨。     

新铌酸盐化合物K4CuNb8O23及其Cu^2＋与Nb^5＋的类质同象代换

1986年Wanklyn等曾报道存在化合物K_5Cu_2Nb_(11)O_(31),并给出其X射线粉晶谱线,但未见进一步指标化和确定其所属晶系及晶胞参数的报道。本研究表明,存在一种新铌酸盐晶体K_4CuNb_8O_(23),其成分及X射线谱异于K_5Cu_2Nb_(11)O_(31)。 用分析纯K_2CO_3、Nb_2O_5及CuSO_4·5H_O按一定比例配制并于1100—1300℃间烧成,     

Effect of nano-CaCO3 on polymorphic behavior in syndiotactic polystyrene for non-isothermal crystallization

Syndiotactic polystyrene (sPS)/nano-CaCO₃ composites were prepared on a twin screw extruder for the first time. Differential scanning calorimeter and wide-angle X-ray diffraction were used to investigate the effect of maximum melting temperature (T_max) and the content of nano-CaCO₃ on the crystal forms and melting behavior of sPS for non-isothermal crystallization. The results indicated that the increasing T_max decreases the crystallization temperature of sPS and facilitates the formation of β-crystal. The addition of nano-CaCO₃ changes the melting behavior and crystal form of melt-crystallized sPS and facilitates the formation of β-crystal. The increase in the content of nano-CaCO₃ decreases crystallization temperature of sPS and increases the content of β-crystal, which was similar to the effect of increasing T_max. Pure β-crystal in sPS/nano-CaCO₃ composites can be obtained at lower melting temperature due to the effect of nano-CaCO₃.     

新型乙烯基聚酯树脂的开发

本文主要介绍了乙烯基聚酯树脂的性能及应用.     

立式聚合釜气液传质特性的研究

本文旨在开发新一代聚四氟乙烯立式反应釜。分别在纯水-氧、乳化剂水溶液-氧两个体系中,用纯氧物理吸收法测定了H/D=1.27的釜的传质容量系数K_La,分析了两体系K_La差别的原因,并对实验结果进行了关联。研究表明,水体系的K_La,是乳液体系的2.5倍,桨型、安装位置等参数对传质有明显影响。在H/D≤1.27的釜中以单层三叶船舶推进器为最佳桨型。     

ACM／FPM并用胶的性能研究

对不同比例的国产ＡＣＭ与ＦＰＭ的并用胶性能和经济成本进行了研究。并用胶保持了良好的物理机械性能和耐高温老化性能，成本明显低于ＦＰＭ，加工性能得到改善，具有广阔的应用前景。     

Noise of He-Cd laser and its suppression

Linear relations have been observed between the reciprocal of the noise ratio and the reciprocal of the output power of the He-Cd laser as well as between the noise ratio and the net gain within the laser cavity. By a feedback method with an acoustooptic modulator in the laser cavity at the Brewster angle for low optical loss, a stable laser with high power output and low noise has been obtained. The noise ratio and the output power are 0.8% and 33 mW, respectively, at 4416 A