广义计算系统研究

该文简要介绍运用广义计算技术而建立起来的一类人工智能系统──广义计算系统，包括广义计算系统的数学描述，广义计算系统的实现问题和学习问题。     

求解PageRank的多步幂法修正的广义二级分裂迭代法

本文基于计算PageRank的广义二级分裂迭代算法,提出了多步幂法修正的广义二级分裂迭代方法.首先,我们详细介绍了该算法的计算过程.然后,证明了该算法的收敛性,并讨论了迭代参数的选取.最后,通过数值实验说明该算法具有比广义二级分裂迭代方法更少的计算开销和更快的收敛速度.     

计算实对称矩阵广义特征值问题的并行算法

矩阵广义特征值问题是科学计算与工程应用中的一个重要的研究课题。文章探讨了近年来计算对称矩阵广义特征值问题的并行算法,并着重介绍了二分法、分治算法、同伦连续法和迭代算法。     

广义系统的反馈控制

本文介绍了广义系统的状态（输出）反馈的一种方便计算方法及反馈作用。讨论了系统的能控性结构指数并指出它是广义系统的不变量。最后研究了广义系统的极点结构配置和静态输出反馈下的极点配置。     

具有DP的广义可能性模糊时态CTL模型检测

为了增强计算树逻辑在时序上的表达能力,以广义可能性测度、决策过程和计算树逻辑为基础,研究了具有决策过程的广义可能性模糊时态计算树逻辑的模型检测。首先采用广义可能性决策过程作为系统模型;然后引入模糊时态算子,构造了模糊时态计算树逻辑并给出其在广义可能性测度下的语义,得到新的广义可能性模糊时态计算树逻辑用来描述系统属性;最后在广义可能性调度下通过模糊矩阵运算讨论了"soon、within、last、nearly"等几类模糊时态连接词的具体计算方法,给出相应的模型检测算法。经验证明,广义可能性模糊时态计算树逻辑是广义可能性计算树逻辑在模糊时序上的扩充,具有更强的表达能力。     

脉冲神经膜计算系统的研究进展及展望(英文)

脉冲神经膜计算系统是膜计算的一个新分支,不仅具有计算机科学研究意义,而且在生物建模和仿真方面有潜在的应用价值．文中介绍了脉冲神经膜计算系统的基本概念并给出了脉冲神经膜计算系统的一个例子;除了介绍标准脉冲神经膜计算系统外,还介绍若干广义的脉冲神经膜计算系统;概述了脉冲神经膜计算系统计算完备性和计算有效性两个方面的主要结果;指出了进一步研究脉冲神经膜计算系统的两个方向．     

基于广义计算的时间序列数据挖掘算法

首先介绍了广义计算和时间序列数据挖掘，然后提出一种基于广义计算的时间序列数据挖掘算法。该算法由灰色回归与神经网络结合而成，通过累加生成降低原始数据的波动幅度，以取得较好的拟合效果，再利用神经网络来完成数据的还原过程。该模型在对浙江省可持续发展指标的预测中，取得了满意的效果。     

挠性结构广义特征值问题的简化处理

本文讨论了挠性飞行器混合坐标动力学模型广义特征值问题的简化计算方法,导出了由三阶行列式计算整体广义特征值和三阶线性代数方程计算整体广义特征向量的公式,所得结果可使得处理高维广义特征值问题大为简化,并可为挠性结构模型降阶提供方便,文中还给出了两个数值例子。     

优势关系下广义决策约简和上近似约简

论文定义了决策表的优势关系下广义决策约简和上近似约简,给出了优势关系下广义决策约简和上近似约简的判定定理和辨识矩阵。同计算优势关系下上近似约简的辨识矩阵相比,计算优势关系下广义决策约简的辨识矩阵的时间复杂度低,由于论文已证明优势关系下广义决策约简和上近似约简是等价的,因此,可以利用优势关系下广义决策约简的辨识矩阵计算优势关系下广义决策约简和上近似约简。     

分布计算与分布控制

本文引述较为广义的概念建立对分布计算和分布控制的统一理解。说明了分布计算和并行计算的共同点和区别。从方法上介绍并行计算的原理,强调了模块间通信量在实现会布计算中的重要影响,并由此出发介绍在分布环境下的任务分配算法。接着还简述了分布控制的概念和结构。然后提出关于分布计算研究现状和发展前途的一些看法。     

Efficiency measurements in heterogeneous distributed computing systems: from theory to practice

This paper addresses the problem of the measurement of efficiency in heterogeneous distributed computing systems. After a discussion on the unsuitability of the traditional notion of efficiency for such systems, a new efficiency metric (generalized efficiency) is introduced by finding the analytical expression of the maximum speedup that can be achieved for a given problem in a heterogeneous system. Generalized efficiency takes into account the heterogeneity of computing resources and is consistent with the definition commonly used for homogeneous systems. The main characteristics of this metric are illustrated, and the problem of performing practical measurements is dealt with thoroughly. Four case studies exploring the whole spectrum of heterogeneous computing systems and of possible classes of applications are presented. Finally, the proposal is compared to related work, and the utility and the limits of generalized efficiency are discussed. © 1998 John Wiley & Sons, Ltd.     

广义信息系统及其决策规则

提出一类更广泛的信息系统,称其为广义信息系统．它包含了完备信息系统、不完备信息系统和多值信息系统．给出了广义信息系统集合近似的概念、方法和相关性质．讨论了广义信息决策系统的决策描述形式．这种决策描述形式易于转化为Skolem标准型，可以直接应用于人工智能的归结推理．     

Diffusion Schemes for Load Balancing on Heterogeneous Networks

Several different diffusion schemes have previously been developed for load balancing on homogeneous processor networks. We generalize existing schemes, in order to deal with heterogeneous networks. Generalized schemes may operate efficiently on networks where each processor can have arbitrary computing power, i.e., the load will be balanced proportionally to these powers. The balancing flow that is calculated by schemes for homogeneous networks is minimal with regard to the l ₂ -norm and we prove this to hold true for generalized schemes, too. We demonstrate the usability of generalized schemes by a number of experiments on several heterogeneous networks.     

广义可能性互模拟及其逻辑刻画

首先,给出了基于广义可能性测度的计算树逻辑的扩展GPoCTL*、计算树逻辑的约简GPoCTL-以及带回报的计算树逻辑GPoRCTL的语构和语义。在经典互模拟和广义可能性测度的基础上讨论了广义可能性互模拟及其相关性质。最后证明了GPoCTL、GPoCTL*和GPoCTL-公式与互模拟状态之间的等价关系。     

Angular momentum in molecular quantum mechanical integral evaluation

Solid-harmonic derivatives of quantum-mechanical integrals over Gaussian transforms of scalar, or radial, atomic basis functions create angular momentum about each center. Generalized Gaunt coefficients limit the amount of cross differentiation for multi-center integrals to ensure that cross differentiation does not affect the total angular momentum. The generalized Gaunt coefficients satisfy a number of other selection rules, which are exploited in a new computer code for computing forces in analytic density-functional theory based on robust and variational fitting of the Kohn-Sham potential. Two-center exponents are defined for four or more solid-harmonic differentiations of matrix elements. Those differentiations can either build up angular momentum about the centers or give forces on molecular potential-energy surfaces, thus generalized Gaunt coefficients of order greater than the number of centers are considered. These 4-j generalized Gaunt coefficients and two-center exponents are used to compute the first derivatives of all integrals involving all the Gaussian exponents on a triplet of centers at once. First all angular factors are contracted with the corresponding part of the linear-combination-of-atomic-orbitals density matrix. This intermediate quantity is then reused for the nuclear attraction integral and the integrals corresponding to each basis function in the analytic fit of the Kohn-Sham potential in the muffin-tin-like, but analytic, Slater-Roothaan method that allows molecules to dissociate into atoms having any desired energy, including the experimental electronic energy. The energy is stationary in all respects and all forces precisely agree with a previous code in tests on small molecules. During geometry optimization of an icosahedral C₇₂₀ fullerene computing these angular factors and transforming them via the 4-j generalized Gaunt coefficient takes more than sixty percent of the total computer time. These same angular factors could be used in identical fashion with Gaussian transforms of Slater-type and numerical radial atomic orbitals.     

基于NPV广义超立方体最佳容错路由算法

在网络可靠性研究中,设计较好的容错路由策略、尽可能多地记录系统中最优通路信息,一直是一项重要的研究工作.超立方体系统的容错路由算法分为可回溯算法和无回溯算法.一般说来,可回溯算法的优点是容错能力强:只要消息的源节点和目的节点有通路,该算法就能够找到把消息传递到目的地的路径;其缺点是在很多情况下传递路径不能按实际存在的最短路径传递.其代表是深度优先搜索(DFS)算法.无回溯算法是近几年人们比较关注的算法.该算法通过记录各邻接节点的故障信息,给路由算法以启发信息,使消息尽可能按实际存在的最短路径传递.这些算法的共同缺点是只能计算出Hamming距离不超过n的路由.在n维超立方体系统连通图中,如果系统存在大量的故障,不少节点对之间的最短路径大于n,因此,这些算法的容错能力差.提出了一个实例说明采用上述算法将遗失60%的路由信息.另外,由于超立方体的结构严格,实际中的真正超立方体系统不多.事实上,不少的网络系统可转换为具有大量错误节点和错误边的超立方体系统.因此,研究能适应具有大量错误节点和错误边的超立方体系统的容错路由算法是一个很有实际价值的工作.研究探讨了:(1) 定义广义超立方体系统;(2) 在超立方体系统中提出了节点通路向量(NPV)概念及其计算规则;(3) 提出了中转点技术,使得求NPV的计算复杂度降低到O(n);(4) 提出了基于NPV的广义超立方体系统最佳容错路由算法(OFTRS),该算法是一种分布式的和基于相邻节点信息的算法.由于NPV记录了超立方体系统全部最优通路和次最优通路的信息,在具有大量故障的情况下,它不会遗漏任何一条最优通路和次最优通路信息,从而实现了高效的容错路由.在这一点上,它优于其他算法.     

Generalized point operators for aggregating intuitionistic fuzzy information

We first develop a series of intuitionistic fuzzy point operators, and then based on the idea of generalized aggregation (Yager RR. Generalized OWA aggregation operators. Fuzzy Optim Decis Making 2004;3:93–107 and Zhao H, Xu ZS, Ni MF, Liu SS. Generalized aggregation operators for intuitionistic fuzzy sets. Int J Intell Syst 2010;25:1–30), we develop various generalized intuitionistic fuzzy point aggregation operators, such as the generalized intuitionistic fuzzy point weighted averaging (GIFPWA) operators, generalized intuitionistic fuzzy point ordered weighted averaging (GIFPOWA) operators, and generalized intuitionistic fuzzy point hybrid averaging (GIFPHA) operators, which can control the certainty degrees of the aggregated arguments with some parameters. Furthermore, we study the properties and special cases of our operators. © 2010 Wiley Periodicals, Inc.     

基于共轭梯度搜索的广义特征对追踪算法

研究广义特征对追踪算法,通过探索基于共轭梯度搜索的标准特征向量追踪算法,将其引入到广义特征对的提取.所提算法具有自适应步长机制,使不同特征搜索方向上的广义瑞利熵达到最优,并适用于提取平稳矩阵束和非平稳矩阵束的广义特征对.数值仿真中将所提算法与多个自适应广义特征向量提取算法进行了比较,实验结果验证了所提算法的有效性.     

Bounds on the generalized μ-scrambling indices of primitive digraphs

Huang and Liu [Generalized scrambling indices of primitive a digraph, Linear Algebra Appl. 433 (2010), pp. 1798–1808] gave the definitions of the generalized μ-scrambling indices and the generalized scrambling indices of a primitive digraph and also provided some exact lower and upper bounds for the generalized scrambling indices of various classes of primitive digraphs. In this paper, we give some exact lower and upper bounds for the generalized μ-scrambling indices of various classes of primitive digraphs.