异构计算系统中弹性节能调度策略研究

目前,节能已成为异构计算系统中减少电量开销、提高系统可靠性和保护环境的重要研究内容.传统的节能调度策略侧重于研究如何节能而忽略了用户对任务完成时间的期望,使得任务执行效果受到较大影响.特别是当系统负载较重时,由于电压调节缺乏自适应性,导致在某些情况下(如应急服务)的任务执行效果不可容忍.文中提出一种弹性节能调度策略(Elastic Energy-Aware Scheduling,EEAS),用于动态调度异构计算系统中非周期、独立任务.EEAS策略根据系统负载情况在系统节能与用户期望之间进行权衡,即当系统负载较重时,EEAS优先考虑用户期望,通过动态调整计算节点局部队列中等待任务的执行电压提高任务完成率;当系统负载较轻时,EEAS在尽量满足用户期望的基础上最大限度地降低任务执行电压以实现节能.文中通过大量的模拟实验比较了EEAS、GEA、HVEA和LVEA的性能.实验结果表明,EEAS的调度质量优于其他策略,可有效提高系统弹性.     

一种新型指数混沌系统及其DSP数字化实现

该文提出了一种基于Chen系统的新型三维指数混沌系统.重点分析了系统参数对整个混沌系统的影响,给出了混沌系统吸引子图,Lyapunov指数图,分叉图等.最后利用DSP实现了该混沌系统,并对由系统生成的数字序列做了NIST测试,测试结果表明序列性能良好.     

EXPERIMENTAL ESTIMATION OF A SCALING EXPONENT FOR TIP VORTEX CAVITATION VIA ITS INCEPTION TEST IN FULL- ANDMODEL-SHIP

Tip vortex cavitation noise of marine propeller became primary concerns to reduce hazardous environmental impacts from commercial ship or to keep the underwater surveillance of naval ships. The investigations of the tip vortex and its induced noise are normally conducted through the model test in a water cavitation tunnel. However the Reynolds number of model-test is much smaller than that of the full-scale, which subsequently results in the difference of tip vortex cavitation inception. Hence, the scaling law between model-and full-scales needs to be identified prior to the prediction and assessment of propeller noise in full scale. From previous researches, it is generally known that the incipient caivtation number of tip vortex can be represented as a power of the Reynolds number. However, the power exponent for scaling, which is the main focus of this research, has not been clearly studied yet. This paper deals with the estimation of scaling exponent based on tip vortex cavitation inception test in both full-and model-scale ships. Acoustical measurements as well as several kind of signal processing technique for an inception criterion suggest the scaling exponent as 0.30. The scaling value proposed in this study shows slight difference to the one of most recent research. Besides, extrapolation of model-ship noise measurement using the proposed one predicts the full-scale noise measurement with an acceptable discrepancy.     

微博社区中用户行为特征及其机理研究

基于微博数据的实证分析表明,微博用户发布微博的数量及时间间隔均服从幂律分布。通过分析微博用户之间的社会网络关系与用户行为之间的联系,发现微博消息数量的分布幂指数与用户互动指数呈反向关系,且随着互动指数的升高,幂指数的下降趋缓。在排除用户之间的评论、转发行为的影响后,个体用户发微博的时间间隔分布幂指数与用户互动指数呈正相关关系。基于上述实证分析,引入用户间评论转发行为、用户间的社会网络关系作为微博用户行为的驱动因素,提出了一种基于社会关系的动力学模型,模型仿真结果与实际数据一致。     

基于ZigBee的动态WSN节点定位

针对因障碍物影响而无法实现动态无线传感网MWSN (mobile wireless sensor network)高精度定位的问题,提出了基于RSSI信号传播模型的α-驱动更新的质心定位算法α- drive renewing weighted centroid localization(α-DRWCL).根据WCL (weighted centroid localization)的距离计算公式,建立了路径衰减因子α的计算模型,利用边界探测装置的探测结果适时改变α的值,得到较接近实际情况的各参考节点权重,达到了减小定位误差的目的.在实验环境中对α-DRWCL算法进行定位仿真,仿真结果表明了该算法的精确性和有效性.     

一种基于顺序多维定标的无锚点定位算法的研究

针对经典多维定标无锚点定位(MDS(C))算法在节点定位精度方面的不足,提出了一种基于顺序多维定标的无锚点定位(MDS(O))算法。该算法假定节点对之间的最短路径距离和欧氏距离存在单调性关系,避免了MDS(C)中节点对之间的最短路径距离和欧氏距离线性关系难以确定的缺陷。理论分析和仿真结果表明:MDS(O)能够提高平均定位精度1.04%～38.2%,有效地实现了节点定位。     

多核系统中基于Global EDF 的在线节能实时调度算法

随着多核系统能耗问题日益突出,在满足时间约束条件下降低系统能耗成为多核实时节能调度研究中亟待解决的问题之一.现有研究成果基于事先已知实时任务属性的假设,而实际应用中,只有当任务到达之后才能够获得其属性.为此,针对一般任务模型,不基于任何先验知识提出一种多核系统中基于Global EDF在线节能硬实时任务调度算法,通过引入速度调节因子,利用松弛时间,结合动态功耗管理和动态电压/频率调节技术,降低多核系统中任务的执行速度,达到实时约束与能耗节余之间的合理折衷.所提出的算法仅在上下文切换和任务完成时进行动态电压/频率调节,计算复杂度小,易于在实时操作系统中实现.实验结果表明,该算法适用于不同类型的片上动态电压/频率调节技术,节能效果始终优于Global EDF算法,最多可节能15％～20％,最少可节能5％～10％.     

A novel approach to update summarization using evolutionary manifold-ranking and spectral clustering

Update summarization is a new challenge in automatic text summarization. Different from the traditional static summarization, it deals with the dynamically evolving document collections of a single topic changing over time, which aims to incrementally deliver salient and novel information to a user who has already read the previous documents. How to have a content selection and linguistic quality control in a temporal context are the two new challenges brought by update summarization. In this paper, we address a novel content selection framework based on evolutionary manifold-ranking and normalized spectral clustering. The proposed evolutionary manifold-ranking aims to capture the temporal characteristics and relay propagation of information in dynamic data stream and user need. This approach tries to keep the summary content to be important, novel and relevant to the topic. Incorporation with normalized spectral clustering is to make summary content have a high coverage for each sub-topic. Ordering sub-topics and selecting sentences are dependent on the rank score from evolutionary manifold-ranking and the proposed redundancy removal strategy with exponent decay. The evaluation results on the update summarization task of Text Analysis Conference (TAC) 2008 demonstrate that our proposed approach is competitive. In the 71 run systems, we receive three top 1 under PYRAMID metrics, ranking 13th in ROUGE-2, 15th in ROUGE-SU4 and 21st in BE.     

一种基于VFD多核系统的硬实时任务节能调度算法

由于芯片功耗不断增加,节能已成为一个亟待解决的重要问题.基于全局异步局域同步(GALS)及电压频率域(VFD)技术的多核处理器计算平台,提出周期性硬实时任务节能调度算法.首先将给定任务集中的实时任务按最差匹配递减(WFD)策略映射到各个计算核上,使各计算核的利用率相对更加均衡,然后利用静态电压?频率调整策略,将每一个VFD内各计算核的共享运行频率降至此VFD中负载最重的计算核的利用率以回收并利用空闲时间节能.在静态策略的基础上提出空闲时间重分配(SR)策略,在保证实时任务可调度的前提下,通过进行任务迁移来平衡VFD内各计算核上的空闲时间分布,以进一步降低VFD的共享运行频率,从而降低能耗.实验表明提出的节能算法可取得较好的节能效果.