区间联想神经网络的稳定性

本文用矩阵测度研究了一类区间联想神经络的稳定性,给出了其稳定的判别准则,推广和改进了文[1-2]的工作。     

Approximation properties of a multilayered feedforward artificial neural network

We prove that an artificial neural network with multiple hidden layers and akth-order sigmoidal response function can be used to approximate any continuous function on any compact subset of a Euclidean space so as to achieve the Jackson rate of approximation. Moreover, if the function to be approximated has an analytic extension, then a nearly geometric rate of approximation can be achieved. We also discuss the problem of approximation of a compact subset of a Euclidean space with such networks with a classical sigmoidal response function.Dedicated to Dr. C.A. Micchelli on the occasion of his fiftieth birthday, December 1992Research supported in part by AFOSR Grant No. 226 113 and by the AvH Foundation.     

求解极小l_1模和极小l_∞模的神经网络

将极小l_１模和极小l_∞模问题同时映射到一个易于电路实现的神经网络模型上［１］，利用该网络，可获得极小ι１模和极小ι∞模问题的实时最优解。     

Carbon nanotubes as building blocks of quantum dots

Quantum dots have been fabricated with single-wall carbon nanotubes (SWCNTs), and their transport properties have been measured at low temperatures. The single-electron transport measurements revealed the artificial atom characteristics with a shell structure and the Zeeman splitting of single particle states. They have been observed with the metallic SWCNT that includes many electrons, in striking contrast to the case of semiconductor artificial atoms that have a few electrons. The unique features in the SWCNT artificial atom are discussed in terms of the energy scales associated with the quantum dot.     

模糊逻辑中的一些问题研究

本文讨论了模糊逻辑中存在的一些问题。特别地，对与、或、非和蕴含作了更深入的讨论。     

大型水利工程多目标综合评价决策支持系统

本文提出了一套以人工智能为基础,关于大型水利工程多目标决策的决策支持系统的框架,该系统集数据库、知识库、决策方法以及选择规则为一体,能较好地反映出决策人的偏好,决策人对于不同的问题,可以根据自己的偏好得到满意的决策结果.     

扫描电镜人工附色摄影技术的研究

本文介绍一种新的彩色摄影法——扫描电镜人工附色摄影技术,此方法利用三原色色补偿滤光片和一枚中黄滤光镜,可以很容易地直接从扫描电镜荧光屏上拍摄到彩色照片。彩色可增加照片的反差对比,增强表现力,在一定程度上弥补了电镜的一些固有缺陷。     

Adaptive station grouping: a highly efficient protocol for optical LANs

A new medium access control protocol for wavelength division multiplexing (WDM) Star Networks, which is capable of achieving a high performance under bursty traffic conditions, is introduced. According to the proposed protocol, the access of stations to the wavelengths is based on their grouping. A separate set of groups is maintained for each wavelength within the traffic stream. All the groups are granted permission to transmit on the corresponding wavelength in a round-robin fashion. The main objective of the grouping algorithm is to have in each group exactly one station which is ready to transmit on the corresponding wavelength, at any given time. In this way, idle slots and collisions are minimized and a nearly optimal throughput-delay performance is achieved. The grouping of stations is dynamically modified at each time slot according to the network feedback information. Due to the dynamic nature of the grouping algorithm, the protocol is capable of being adapted to the sharp changes of the stations’ traffic. Extensive simulation results are presented which indicate that the proposed protocol achieves a significantly superior performance than other well-known medium access control protocols for WDM Star Networks.