基于“投资-效益”准则的RC桥墩全寿命总费用

基于性能的抗震设计于20世纪90年代兴起以来,目前在土木建筑方面的文献较多,但是关于桥梁的研究成果较少。采用五级钢筋混凝土桥墩结构性能水平,以钢筋混凝土桥墩墩顶位移角为量化指标,建立了钢筋混凝土桥墩性能水平与钢筋混凝土桥墩位移角限值之间一一对应的关系。将地震危险性分析与最优设防烈度结合起来,建立基于投资-效益准则的RC桥墩全寿命总造价计算的通用框架。     

Dual–primal proximal point algorithms for extended convex programming

This paper presents a class of dual–primal proximal point algorithms (PPAs) for extended convex programming with linear constraints. By choosing appropriate proximal regularization matrices, the application of the general PPA to the equivalent variational inequality of the extended convex programming with linear constraints can result in easy proximal subproblems. In theory, the sequence generated by the general PPA may fail to converge since the proximal regularization matrix is asymmetric sometimes. So we construct descent directions derived from the solution obtained by the general PPA. Different step lengths and descent directions are chosen with the negligible additional computational load. The global convergence of the new algorithms is proved easily based on the fact that the sequences generated are Fejér monotone. Furthermore, we provide a simple proof for the O(1/t) convergence rate of these algorithms.     

数据中心网络TCP Incast仿真分析

为分析云计算数据中心网络中的TCP Incast问题的产生原因并探讨缓解该问题的有效途径,采用基于NS2仿真实验的方法,通过改变TCP协议类型、交换机缓冲区大小、服务请求单元大小及重传超时RTO值等影响因素,分析传输吞吐率性能的变化。分析结果表明,通过改变这些影响因素有助于缓解TCP Incast问题,但同时也存在各种缺点和局限性。     

Point-augmented biquadratic C1 subdivision surfaces

Shape artifacts, especially for convex input polyhedra, make Doo and Sabin’s generalization of bi-quadratic (bi-2) subdivision surfaces unattractive for general design. Rather than tuning the eigenstructure of the subdivision matrix, we improve shape by adding a point and enriching the refinement rules. Adding a guiding point can also yield a polar bi-2 subdivision algorithm. Both the augmented and the polar bi-2 subdivision are complemented by a new Primal Bi-2 Subdivision scheme. All surfaces are $C^1$ and can be combined.     

Efficient digital certificate verification in wireless public key infrastructure using enhanced certificate revocation list

ABSTRACT

Public key cryptography is widely used for secure data transfer in various applications. The Public key infrastructure provides a foundation that supports public key cryptography, and it is required to deliver the public keys to existing systems or users securely. The public key is exchanged digitally in the form of digital certificates having a certain period of validity. Circumstances may arise under which the validity of the certificates should be ceased, and they need to be revoked. The fast growth of e-commerce demands verification and data transfers to be done in minimum time to maintain high efficiency of communication. In this paper, a new method of developing certificate revocation lists is proposed. The formal model of enhanced PKI for mobile commerce security is put forward. The performance of NTRU is compared with the RSA algorithm and the ElGamal algorithm. The paper also sets forth a comparative analysis of the proposed method with other existing techniques. The results show that the proposed technology has the upper hand over the current methods.     

Virtual life: self and identity redefined in the new media age

In this article, the author discusses her virtual reality installation The Parallel Dimension and goes on to explore the conflicts of identity experienced when a person is confronted with VR as a total immersion. She suggests that the various questions arising through this experience may lead us to a more ambivalent concept of the self.     

基于HCM的可扩展路由器体系结构MER

本文提出了基于层次化协同模型的多维可扩展路由器MER。参考Open Router路由器模型,利用层次化协同模型实现了独立于数据层的控制平台CSP,奠定了开放控制的基础;基于功能插件和抽象插件构造了基于插件的路由器抽象层PRA,分离了控制与转发平面,支持路由器转发平面功能的动态扩展;并举例说明了MER中高效的网络应用控制路由选择机制,体现了直接控制原理,有助于设计更为复杂的控制策略。模拟结果表明,MER具有良好的扩展性,支持网络应用直接控制,为控制平面智能化奠定基础。     

基于计算缓存方法的分子动力学程序性能优化

分子动力学数值模拟程序在现代高性能计算机上的计算效率往往很低,只能发挥系统峰值性能的几个百分点。本文对并行分子动力学程序PMD3D在联想深腾6800超级计算机上进行性能优化。通过性能分析,我们发现粒子相互作用力计算中相互关联的浮点运算严重影响了处理器的指令级并行效率,为此我们应用计算缓存的方法,将大量不规则的浮点计算进行缓存,达到一定规模后再进行向量化计算。这样使得单机性能在优化后提升4倍多,达到处理器峰值性能5.2GFlops的32.3%。最后,在深腾6800的64个节点的256个CPU上进行了并行性能测试,达到峰值运算性能1.3万亿次的27%。     

高性能计算中处理器功耗特征的评测与分析

高性能计算系统的系统结构和应用模式与单机系统或商用机群服务器系统都有很大的不同,掌握功耗特征是提高能效的前提。本文将支撑功耗管理的低功耗技术分为动态资源休眠和动态速率调节两类,并就处理器的这两类机制在高性能计算中的应用进行评测,验证了功耗管理在高性能计算中的有效性,量化分析了处理器功耗特征,指出了当前管理方案的不足及改进设想,对进一步能耗优化有重要的指导意义。     

改进自组织映射神经网络在指纹识别中的应用

改进自组织映射神经网络方法是将常规自组织映射神经网络方法结合确定性水平,对网络的输入矢量进行预处理。通过实验比较了这种改进的自组织映射神经网络识别方法与常规的自组织映射神经网络识别方法的识别效果,在识别性能上有了很大的提高。