基于DIGNET网络的数据融合方法

针对数据融合和目标识别的特点,提出了基于DIGNET自组织聚类人工神经网络的数据融合方法。考虑到多传感器系统测量多个参量的特点,用并行的子网络结构代替中间隐层,实现了基于决策层的信息融合目标识别。利用仿真数据对基于DIGNET的数据融合方法进行了实验研究。实验结果表明,该方法具有数据正确分类率高和抗噪能力强等优点,有效地实现了融合识别。将该方法应用于前视红外和可见光双传感器目标跟踪系统的数据融合识别是可行的。     

考虑联盟结构的联合采购合作剩余分配博弈分析

联合采购合作剩余分配本质是在最大联盟剩余值与最小期望分配值之间的谈判协商过程。当前联合采购合作剩余分配研究主要基于参与企业地位无差别假设,忽略了联盟结构对剩余分配具有明显影响。基于联盟形成的层次结构,研究论证了采购企业通过组建小联盟形式参与更大联盟的联合采购,可提升其在分配谈判中的话语权。通过算例分析表明,考虑联盟结构的联合采购合作剩余分配相对于不考虑联盟结构得到的分配方案更具公平性和合理性。     

HL-2A tokamak disruption forecasting based on an artificial neural network

Artificial neural networks are trained to forecast the plasma disruption in HL-2A tokamak. Optimized network architecture is obtained. Saliency analysis is made to assess the relative importance of different diagnostic signals as network input. The trained networks can successfully detect the disruptive pulses of HL-2A tokamak. The results obtained show the possibility of developing a neural network predictor that intervenes well in advance for avoiding plasma disruption or mitigating its effects.     

计算神经科学

计算神经科学是近三十年来出现的一个新兴交叉学科,它强调采用数学定量的方法,如数学建模、理论分析和数值模拟等来研究和解决神经科学中的重要科学问题,一方面神经科学实验现象为发展新的数学模型、理论和算法提供了基础,另一方面通过数学定量,能反过来揭示神经科学实验现象背后的数理机制、发现新的科学规律.随着欧盟、美国、日本和我国脑...     

The Detection of Defects in Optical Fibers Using a Hybrid Opto-electronic Correlator

A hybrid opto-electronic correlator for detecting defects in optical fibers is proposed. After the light from a He-Ne laser being expanded and filtered it is not collimated but directly passes a Fourier transform lens and illuminates a reference fiber and a test fiber at the same input plane. The Fourier transform spectrum of the two fibers is therefore obtained at the rear focal plane of the lens, where it is sampled via a CCD array connected with a computer through a frame grabber. The computer performs filter, inverse Fourier transform and setting threshold operation on classification. The system is an equivalent of joint transform correlator with a Fourier lens of long focal length. The experiment results for optical fibers having incoordinate defects are presented. The results indicate that the system can be used for fiber defect detection, and has the advantages of high identification, compact configuration, easy adjustment and flexible manipulation.     

宏观调控“两难”新环境下中国CPI运行定量分析——基于改进的BP神经网络

当前,我国经济继续回升向好,但宏观调控面临的"两难"问题增多.如何防止经济快速下滑和防范通货膨胀成为当前宏观经济决策的重要内容,准确预见我国价格运行趋势对宏观决策具有十分重要的意义.1997年和2008年分别发生了对世界经济发展影响深远的两次金融危机.两次金融危机发生后,中国政府均实施了大规模的经济刺激计划,但反映经济运行的主要宏观经济指标CPI月度指数从危机发生当年的次年开始,步入了长达数月负区间运行.研究发现1997年金融危机发生前后月度CPI指数运行相关度非常高,2008年金融危机发生当年月度CPI指数与2007年、2009年月度CPI指数运行相关度也非常高.对此,运用改进的BP神经网络模型对1995-1999年中国月度CPI指数进行学习,以2006-2009年月度CPI指数数据为基础,对2010年中国下半年月度CPI指数进行了预测,显示全年CPI指数预计为102.85.对此,认为2010年我国面临适度的通胀压力,提出了应对政策建议.     

A Markov Model of Production, Trade, and Money: Theory and Artificial Life Simulation

The paper generalizes the Kiyotaki-Wright trade model by treatingthe trading period as a finite game, so Nash's theorem can be used to provethe existence of equilibrium, and by treating the economy as a Markovprocess, so an ergodic theorem can be used to show the existence ofequilibria with desirable properties (e.g., in which money exists). A Markovmodel of trade also allows us to add complexity to the economy withoutadding corresponding complexity to the analysis of the model's properties.The paper also provides artificial life simulations of the Markov economysuggesting that monetary equilibria are dynamically stable and do notrequire high levels of learning or information processing on the part ofagents.     

用联合变换相关器实现灰阶图像绝对差度量光学匹配

将绝对差度量（ＡＤＭ）引入联合变换相关器（ＪＴＣ）,实现了灰阶图像光学匹配的计算机模拟实验,利用循环编码方法,通过联合变换相关器一步绝对差度量,相关后的图像通过取高阈值的分割方法来实现对目标的准确判断,针对高斯噪声干扰情况,分析了阈值选择问题,提高了灰阶目标识别抵抗器材怕畸变干扰的能力。     

一种基于功率谱对数变换的联合变换相关器

联合功率谱的高频成份对相关峰识别具有重要作用,实验中记录的联合功率谱随着频率的增加其强度迅速降低。为增强联合功率谱的高频成分,改善其调制程度,可对功率谱进行对数变换,从而有效保留更多的高频成份,压低零级衍射光斑,提高联合变换相关器的识别能力。     

Biophysical changes induced by cholesterol on phosphatidylcholine artificial biomembranes containing alamethicin oligomers

Cholesterol is an important constituent of eukaryotic cell membranes, whose interaction with phospholipids leads to a broad range of biological roles, such as: maintenance of proper fluidity, formation of raft domains, reduction of passive permeability of various chemical species through the bilayer (e.g., glucose, glycerol, K⁺, Na⁺ and Cl⁻ ions), and increased mechanical strength of the membrane. In this work we studied an interesting paradigm, as to whether cholesterol-containing phosphatidylcholine biomembranes influence the kinetics and transport features of alamethicin oligomers embedded into it. We demonstrate that moderate relative amounts of cholesterol increase the electrical conductance of various sub-conductance states of the alamethicin oligomer, caused probably by a non-monotonic change in the lumped dipole moment of the biomembrane. Our data suggest that biomembrane stiffness caused by cholesterol, visibly modifies the association-dissociation rates of alamethicin oligomerization in the biomembrane. Moreover, increasing concentrations of cholesterol seem to lead to more stable intermediate alamethicin oligomers. We show that in the presence of cholesterol, as the diameter of the alamethicin oligomer increases, so does the time of another monomer to get picked up. These results brings into focus the interesting issue of how oligomerization of proteins affects their interaction affinities for membrane-based lipids.