Dynamic file grouping for load balancing in streaming media clustered server systems

A dynamic file grouping strategy is presented to address the load balancing problem in streaming media clustered server systems. This strategy increases the server cluster availability by balancing the workloads among the servers within a cluster. Additionally, it improves the access hit ratio of cached files in delivery servers to alleviate the limitation of I/O bandwidth of storage node. First, the load balancing problem is formulated as a two layers semi-Markov switching state-space control process. This analytic model captures the behaviors of streaming media clustered server systems accurately, and is with constructional flexibility and scalability. Then, a policy iteration based reinforcement learning algorithm is proposed to optimize the file grouping policy online. By utilizing the features of the event-driven policy, the proposed optimization algorithm is adaptive and with less computational cost. Simulation results demonstrate the effectiveness of the proposed approach. Recommended by Editor Hyun Seok Yang. This work was supported by the National Natural Science Foundation of China under grant Nos. 60774038, 60574065, National 863 HI-TECH Research & Development Plan of China under grant Nos. 2006AA01Z114, 2008AA01A317, Natural Science Foundation of Anhui Province under grant No. 070412063, Graduate Student Innovation Foundation of USTC under grant No. KD2006036, and Science Research Development Foundation of HFUT under grant No. GDBJ2008-045. Qi Jiang received the B.S. degree in Industrial Electrical Automation from Southeast University in 1989 and the Ph.D. degree in Control Science and Engineering from University of Science and Technology of China in 2008. He is currently a Post-doc in USTC. His research interests include optimization and control of stochastic dynamic systems, and performance analysis and optimization of network communication systems. Hong-Sheng Xi received the M.S. degree in Applied Mathematics from University of Science and Technology of China in 1977. He is currently a Professor in Department of Automation, USTC. His research interests include discrete event dynamic systems, performance analysis and optimization of network communication systems, robust control, and network security. Bao-Qun Yin received the B.S. degree in Mathematics from Sichuan University in 1985, the M.S. degree in Applied Mathematics and the Ph.D. degree in Pattern Recognition and Intelligent Systems from University of Science and Technology of China in 1993 and 1998, respectively. He is currently a Professor in Department of Automation, USTC. His research interests include discrete event dynamic systems, and Markov decision processes.     

Cyclic moving average control approach to cylinder pressure and its experimental validation

Cyclic variability is a factor adversely affecting engine performance. In this paper a cyclic moving average regulation approach to cylinder pressure at top dead center （TDC） is proposed, where the ignition time is adopted as the control input. The dynamics from ignition time to the moving average index is described by ARMA model. With this model, a one-step ahead prediction-based minimum variance controller （MVC） is developed for regulation. The performance of the proposed controller is illustrated by experiments with a commercial car engine and experimental results show that the controller has a reliable effect on index regulation when the engine works under different fuel injection strategies, load changing and throttle opening disturbance.     

基于分布式入侵检测系统的负载均衡算法的比较

随着网络带宽的增长,分布式NIDS逐渐成为入侵检测系统的主流,而其中的负载均衡技术是提高系统性能的关键技术之一。通过模拟的方法,利用真实的网络数据,将多种负载均衡算法分为无连接约束算法和有连接约束算法进行了比较、评价,为分布式入侵检测系统的设计提供负载均衡算法的选择依据。     

异构结构化P2P网络负载均衡方案

结构化P2P网络由于采用DHT算法导致节点存储资源的不均衡,当前解决方案都是假定节点容量及负载是均匀分布在系统中,而忽略了实际网络存在的节点异构性的影响.本文提出的考虑节点异构性的结构化P2P网络负载均衡方案提出了负载均衡的衡量标准--负载平滑度,采用基于相同资源描述符的资源整体转移方案,以节点的邻居节点为平衡范围,描述了系统在节点加入、离开,资源加入以及节点过载情况下的算法,使得整个系统逐步达到负载均衡.该方案充分考虑了实际网络中存在的异构问题.仿真实验表明,该方案有效地解决了并构P2P网络下的负载均衡问题.     

异构无线网络中基于效用的呼叫接纳控制及流量均衡策略

目前国内外对于异构无线网络中联合呼叫接纳控制和流量均衡策略的研究大多停留在网络层以上,即认为不同的网络具有不同的覆盖面积和网络容量,而没有考虑不同网络在物理层和链路层的具体特性.本文在对UMTS和WLAN底层特性分析的前提下,结合应用层为用户设定的效用函数,利用跨层信息提出了一种新的联合呼叫接纳控制及流量均衡策略.仿真结果表明该策略可以使异构无线网络中所有呼叫的效用总和最大.     

Web检索查询意图分类技术综述

查询分类是近年来信息检索领域的研究热点,并且在很多领域得到了广泛地关注。主要讨论根据查询的意图进行分类的研究工作,从查询分类的诞生背景、关键技术、所使用的分类方法和评价方法方面进行综述评论,提出了查询意图分类面临的问题和挑战。认为缺乏权威的评测标准、在大规模数据集上的未经全面测试的性能、如何准确地获取查询的特征以及如何证明分类体系的完备性和独立性是目前查询意图分类研究的关键问题。     

基于Delaunay三角网的空间离群挖掘

空间离群是指非空间属性与其空间邻居显著不同的空间对象。空间数据的特殊性决定了空间离群挖掘需要充分考虑空间数据的特点,才能挖掘出有现实意义的离群。本文对现有主要的空间数据离群挖掘算法进行了研究分析,针对k-邻域法确定空间邻域的缺点,基于Delaunay三角网在表达空间邻近关系的有效性,通过构建Delaunay三角网确定空间邻域并生成空间权重矩阵,据此提出了基于Delaunay三角网的空间离群挖掘算法DT_SOF,并以实际生态地球化学数据进行实验检验。结果表明,算法具有较低的用户依赖性,能准确挖掘空间离群。     

基于共享库的数据集成方案改进

基于共享数据库的方案是现在高校进行数据集成普遍采用的模式。但随着全局业务的增多,这种方案不可避免地会遇到共享数据库性能瓶颈的问题。从有效利用业务数据库、降低共享数据库业务压力出发,在研究了已有的改进方案之后,对基于共享数据库方案的框架进行改进,提出了一个同时使用共享数据库模式和中介模式,能够进行负载均衡的新的数据集成框架模型。对模型的优劣性进行了分析。     

基于复小波变换的遥感图像并行融合算法

随着遥感技术的快速发展,多源遥感图像的快速融合成为很多遥感任务的关键处理步骤。为了加速遥感图像的融合处理过程,本文首先提出了一种新的基于双树复小波变换的并行融合算法（PACWT）。算法中综合运用了数据分布、并行数据处理和负载均衡技术,以克服单机处理在计算能力和存储器空间上的限制;针对基于CWT的图像融合处理的计
计算特点,设计了一种可有效避免计算过程中数据通信的冗余划分方法。然后,从理论上分析了算法在时间和空间两方面的性能。最后,通过实验分析了算法在32-CPU的Cluster系统上的实际性能。结果表明,本文提出的算法具有良好的可扩展性,在数据量较大时可获得良好的加速比和并行效率。     

基于子模块电容电压允许值的MMC均压控制策略

针对模块化多电平变换器（modular multilevel converter,MMC）传统子模块（sub-module,SM）均压策略存在的开关频率高、开关损耗量多、计算量大等问题,提出基于子模块电容电压允许值的均压控制策略。通过引入电容电压允许值变量,使得SM电压始终在规定的范围内运行,降低了功率器件损耗。与此同时,由于不需要每个控制周期对全部SM进行排序,进一步减少了控制策略的计算量。通过仿真结果验证了所提策略的有效性,可以显著减少单位时间内功率器件的开关次数。