Combining Global and Local Surrogate Models to Accelerate Evolutionary Optimization

In this paper, we present a novel surrogate-assisted evolutionary optimization framework for solving computationally expensive problems. The proposed framework uses computationally cheap hierarchical surrogate models constructed through online learning to replace the exact computationally expensive objective functions during evolutionary search. At the first level, the framework employs a data-parallel Gaussian process based global surrogate model to filter the evolutionary algorithm (EA) population of promising individuals. Subsequently, these potential individuals undergo a memetic search in the form of Lamarckian learning at the second level. The Lamarckian evolution involves a trust-region enabled gradient-based search strategy that employs radial basis function local surrogate models to accelerate convergence. Numerical results are presented on a series of benchmark test functions and on an aerodynamic shape design problem. The results obtained suggest that the proposed optimization framework converges to good designs on a limited computational budget. Furthermore, it is shown that the new algorithm gives significant savings in computational cost when compared to the traditional evolutionary algorithm and other surrogate assisted optimization frameworks     

浅析面向业务和客户的IP网管系统

从分析IP网管的现状和所面临的问题入手,简要地分析了面向业务和客户网管系统的需求,并提出了一些相关的设计思路。     

Reliability Models of NMR Systems

Majority voted redundant systems are widely used. A reliability model is developed and analyzed for N-tuple Modular Redundancy-NMR: (n + 1)-out-of-(2n + 1)-where the units are subject to stuck-at-0, stuck-at-1 or stuck-at-X failures and where failures can occur in a mutually compensatory manner. A reconfiguration of the NMR redundancy, the NMR/Simplex strategy, is proposed and evaluated and its model shown to be included in the general model for the compensated NMR.     

Reduced Models of Arterial Systems

Simple models that manifest input impedances of arterial systems are compared. An improvement upon documented two-, three-, and five-element models is presented. The classical two-element model (the windkessel) accounts for the lowest frequency components, and the three-element model (the modified windkessel) accounts for both low-and high-frequency components of the spectrum of interest. Five-element models, however, by allowing for reflection, can account for principal features over the entire frequency range of interest.     

Local and Global Properties of a Lead-Free Solder

Elastic and viscous properties including Young’s modulus, hardness, creep rate sensitivity, and fatigue resistance of Sn-1.2Ag-0.5Cu-0.05Ni lead-free solder have been investigated. The properties of bulk specimens and in situ solder balls are compared. Experiments show good correlations of Young’s modulus and creep rate sensitivity between conventional measurements and nanoindentation results on bulk specimens. Further mechanical properties of the beach-ball microstructure in solder balls are characterized by nanoindentation. The load–partial unload technique has been used to determine the variation in mechanical properties with increasing depth of penetration into the intermetallic inclusions in the in situ solder. The fatigue resistances of the bulk specimens and solder balls are compared by using the novel nanoimpact method. In comparison with bulk specimens, it is found that in situ solder has higher Young’s modulus, lower creep strain rate sensitivity, and better fatigue resistance. The effects of soldering and the scale differences strongly affect the mechanical and fatigue properties of in situ solder.     

Design and Performance Modeling for Single-Walled Carbon Nanotubes as Local, Semiglobal, and Global Interconnects in Gigascale Integrated Systems

Based on physical models, distributed circuit models are presented for single-walled carbon nanotubes (SWCNs) and SWCN bundles that are valid for all voltages and lengths. These models can be used for circuit simulations and compact modeling. It is demonstrated that by customizing SWCN interconnects at the local, semiglobal, and global levels, several major challenges facing gigascale integrated systems can potentially be addressed. For local interconnects, monolayer or multilayer SWCN interconnects can offer up to 50% reduction in capacitance and power dissipation with up to 20% improvement in latency if they are short enough (<20 mum). For semiglobal interconnects, either latency or power dissipation can be substantially improved if bundles of SWCNs are used. The improvements increase as the cross-sectional dimensions scale down. For global interconnects, bandwidth density can be improved by 40% if there is at least one metallic SWCN per 3-nm² cross-sectional area     

面向客户和业务的OSS的研究

本文从面向客户和业务的角度论述和分析了当前形势下OSS(运营支撑系统)的需求、框架体系及实施策略,并介绍了两种典型的面向客户和业务的OSS,指出其在电信运营企业转型过程中的重要支撑作用.     

Taxi dispatching via Global Positioning Systems

This paper empirically examines taxi dispatch operations via Global Positioning Systems (GPS) by three taxi companies. The implementation of GPS technology has radically changed the traditional approach to taxi fleet management. In comparison to the radio-paging systems used previously, the satellite-based dispatch systems have considerably enhanced the accuracy, communications, and productivity of taxi operations. However, variables associated with human-system interfaces and operational environments should be taken into account in order to achieve the full capacity of the systems. Furthermore, the reengineering of operations and the training of end-users are vital to the improvement of service quality and the development of sophisticated public transport systems     

全球卫星导航系统的发展概况

今天,卫星导航系统已经在大量应用中广泛使用,而且总的发展趋势是为实时应用提供高精度服务。迄今为止,比较完善的卫星导航系统已经有美国ＧＰＳ和俄罗斯ＧＬＯＮＡＳＳ系统,它们的定位和定时精度都能满足严格的军用定位和定量要求。而对大量的卫星导航需求,欧洲计划推出自己的卫星导航系统Ｇａｌｉｌｅｏ。本文对现有的ＧＰＳ和ＧＬＯＮＡＳＳ系统进行了描述和比较,并对Ｇａｌｉｌｅｏ系统的计划方案进行了简要说明。     

网格资源管理系统模型研究

网格资源管理系统是网格计算系统的重要组成部分。文章分析了网格资源管理系统的功能需求，具体论述了该模型的原型系统，最后介绍了相关研究工作。     

电力变压器模型和数据采集一体化系统

电力变压器故障会给电力公司造成巨大的损失，因为故障的出现需要电力公司消耗资源来对电网进行恢复并执行周期性的日常维护。为了避免在四个40MVA的电力变压器中出现该问题，作者实现了一个带故障预警功能的监测系统，该系统也是预警维护系统底层网络的基础。通过数据库技术，作者用一种称为虚拟数据卡（VAC）的方法给模型提供经过提炼的传感器采集数据。ACSs和系统模块化结构使得该系统是一种可升级的、可重新配置的和容易维护的系统。     

Clustered Local Decoding for Wyner-Type Cellular Models

We study the uplink of Wyner-type cellular models featuring short range inter-cell interference. We assume that the decoding of the message sent by a given transmitter is done locally, that is using only the signals received at the antennas in its vicinity. We derive upper and lower bounds on the achievable rate for a two-dimensional model and for a one-dimensional model with an attenuation parameter. The analysis of our cellular models will rely on rate splitting and genie-aided bounds, originally employed in the framework of the interference channel.      

一种基于全局和局部标记相关性的多标记分类算法

多标记学习用于处理一个示例同时与多个类别标记相关的问题.在多标记学习中,标记相关性能够显著提升学习算法的性能.大多数现有的多标记学习算法在利用标记的相关性时,要么只使用被所有示例所共享的全局标记相关性,要么就使用局部标记相关性,它们认为不同簇中的示例应该存在不同的标记相关性.本文中,我们提出了一种同时利用全局和局部标记相关性的多标记学习算法,从而为学习进程提供更全面的标记信息.在计算全局和局部标记相关性时,我们使用了余弦相似性来获取不同标记之间的正相关性和负相关性,这样有助于我们进一步实现更可靠的多标记学习.我们在多种类型的数据集上进行了广泛的对比实验来验证所提算法的有效性.实验结果表明,该算法显著优于大多数对比算法,展现出其在多标记学习中的突出性能.     

基于全局和局部保持的半监督支持向量机

 支持向量机（SVM）作为正则化方法的一个特例在模式识别领域得到了成功地运用,然而传统的SVM方法作为一种有监督的学习方法主要依据最大间隔原则得到决策超平面的法向量,而并没有充分考虑样本内在的几何结构以及所蕴含的判别信息. 因此,本文将线性判别分析（LDA）的类内散度和保局投影（LPP）的基本原理引入到SVM中,提出基于全局和局部保持的半监督支持向量机：GLSSVM,该方法在继承传统的SVM方法的特点的基础上,充分考虑样本间具有的全局和局部几何结构,体现样本间所蕴含的局部和全局判别信息,同时满足作为半监督方法的必须依据的一致性假设,从而在一定程度上提高了分类精度．通过在人造数据集和真实数据集上的测试表明该方法具有上述优势．     

Service-Oriented Semantic Interoperability Technologies and Standards

Service-Oriented Software Engineering (SOSE) presents new challenges; in particular, how to promote interoperability and cooperation among loosely-coupled service resources. This is critical for service resource sharing and for implementing on-demand services. This paper discusses key technologies of service virtualization, including encapsulation of service interoperability (for available resources); ontology-based Role, Goal, Process, and Service (RGPS) metamodeling (for interoperable aggregation and organization of virtualization services); registration and repository management of Metamodel Framework for Interoperability (MFI) (for virtualization service management); and virtualization service ontology and its represented association with RGPS. Latest progress of the MFI and ISO standards is also discussed.     

面向服务体系结构的连接器规约及其验证

连接器建模是软件体系结构设计中研究的重要问题之一.同样,考虑到面向服务体系结构松散耦合的特性,对其所涉及的连接独立于服务的设计进行规约也是一个关键问题.本文将SOA中的服务连接抽象为面向服务设计中的第一类实体--面向服务体系结构的连接器.给出了SOA的连接器模型的规约,并采用π演算作为文中方法的形式化基础以实现对其特性...     

基于融合面向业务--ZTE Softswitch系统总体介绍

中兴通讯的ZTE Softswitch系统不仅为运营商构建了一个先进的、可以无限扩展的网络平台，使多种网络向数据网的不断融合成为可能，更重要的是其以业务快速、有效、灵活提供为最终目标，推出多种贴近用户需要的差异化业务，为最终实现多媒体综合通信业务正做出不懈的努力。     

A Global Sensitivity Analysis of Three- and Four-Layer EEG Conductivity Models

The accuracy of forward models for EEG partly depends on the conductivity values of the head tissues. Yet, the influence of the conductivities on the model output is still not well understood. In this paper, we apply a variance-based sensitivity analysis method to the most common EEG forward models (three or four layers). This method is global because it quantifies the influence of each parameter with all the parameters varying at the same time. With nonlinear models, it helps to understand the interaction between parameters, which is not possible with simple sensitivity analyses (one-at-a-time variations, derivatives, and perturbations). By analyzing the potential topographies at the electrodes, we obtained several results. For a shallow dipole, the EEG topographies are mainly sensitive to the interaction between skull and scalp conductivities. It means that the variability of the EEG topographies is driven mostly by a function of skull and scalp conductivities. Similar results are presented for skull anisotropy and a current injection as performed in electrical impedance tomography. This global sensitivity analysis gives new information about EEG forward models—it identifies the main input parameters that need model refinement—and directions on how to calibrate these models.      

Stochastic Models for Multistage Cell Classification Systems

Probabilistic models for multistage cell classification systems are described. A simple finite Markov chain models classification events which occur as a cell passes through the system. The state space consists of various identities assigned to the cell, including true celi type and identities assigned by classifiers. Effects of throughput rate, data buffer capacity, and classifier processing rate on system performance are predicted by another model composed of a network of single server queues. Markov and queue models are interrelated in that classification events at one processor (modeled by the Markov chain) govern arrival rates of other processors. In turn, the queue model predicts the probability that a cell wili be missed due to fmite data buffer capacity. The miss event is modeled by the Markov chain as a possible classification outcome. Application of the models is illustrated for a multistage gynecologic flow prescreening system with slit-scan processing in the first stage and two dimensional image processing in the second. Results predict system sensitivity as a function of first stage false alann rate and abnormal cell occurrence rate.