An Approach for Network Function Combination Based on Least Busy Placement Algorithm

Recently, integrating Softwaredefined networking(SDN) and network functions virtualization(NFV) are proposed to address the issue that difficulty and cost of hardwarebased and proprietary middleboxes management. However, it lacks of a framework that orchestrates network functions to service chain in the network cooperatively. In this paper, we propose a function combination framework that can dynamically adapt the network based on the integration NFV and SDN. There are two main contributions in this paper. First, the function combination framework based on the integration of SDN and NFV is proposed to address the function combination issue, including the architecture of Service Deliver Network, the port types representing traffic directions and the explanation of terms. Second, we formulate the issue of load balance of function combination as the model minimizing the standard deviations of all servers' loads and satisfying the demand of performance and limit of resource. The least busy placement algorithm is introduced to approach optimal solution of the problem. Finally, experimental results demonstrate that the proposed method can combine functions in an efficient and scalable way and ensure the load balance of the network.

Aperiodic Checkpoint Placement Algorithms—Survey and Comparison

In this article we summarize some aperiodic checkpoint placement algorithms for a software system over infinite and finite operation time horizons, and compare them in terms of computational accuracy. The underlying problem is formulated as the maximization of steady-state system availability and is to determine the optimal aperiodic checkpoint sequence. We present two exact computation algorithms in both forward and backward manners and two approximate ones;constant hazard approximation and fluid approximation, toward this end. In numerical examples with Weibull system failure time distribution, it is shown that the combined algorithm with the fluid approximation can calculate effectively the exact solutions on the optimal aperiodic checkpoint sequence.

Repulsive firefly algorithm-based optimal switching device placement in power distribution systems

To achieve optimal configuration of switching devices in a power distribution system,this paper proposes a repulsive firefly algorithm-based optimal switching device placement method.In this method,the influence of territorial repulsion during firefly courtship is considered.The algorithm is practically applied to optimize the position and quantity of switching devices,while avoiding its convergence to the local optimal solution.The experimental simulation results have showed that the proposed repulsive firefly algorithm is feasible and effective,with satisfying global search capability and convergence speed,holding potential applications in setting value calculation of relay protection and distribution network automation control.

A NOVEL 2-D GENETIC ALGORITHM FOR STANDARD PE PLACEMENT IN OPTOELECTRONIC MCM

In order for optical interconnection technologies to be incorporated into the next generation parallel computers, new optoelectronic computer aided design, integration, and packaging technologies must be investigated. One of the key issues in designing is the system volume, which is determined by maximum interconnection distance(MID) between PEs. A novel 2 D genetic algorithm was presented in this paper at the first time, and used to solve the placement of twin butterfly multistage networks based on transmissive physical model. The experiment result shows that this algorithm case works better than other algorithm cases.

A Data-Placement Strategy Based on Genetic Algorithm in Cloud Computing

With the development of Computerized Business Application, the amount of data is increasing exponentially. Cloud computing provides high performance computing resources and mass storage resources for massive data processing. In distributed cloud computing systems, data intensive computing can lead to data scheduling between data centers. Reasonable data placement can reduce data scheduling between the data centers effectively, and improve the data acquisition efficiency of users. In this paper, the mathematical model of data scheduling between data centers is built. By means of the global optimization ability of the genetic algorithm, generational evolution produces better approximate solution, and gets the best approximation of the data placement at last. The experimental results show that genetic algorithm can effectively work out the approximate optimal data placement, and minimize data scheduling between data centers.

Algorithms for the Optimization of Well Placements—A Comparative Study

The Artificial Bee Colony (ABC) is one of the numerous stochastic algorithms for optimization that has been written for solving constrained and unconstrained optimization problems. This novel optimization algorithm is very efficient and as promising as it is;it can be favourably compared to other optimization algorithms and in some cases, it has been proven to be better than some known algorithms (like Particle Swarm Optimization (PSO)), especially when used in Well placement optimization problems that can be encountered in the Petroleum industry. In this paper, the ABC algorithm has been modified to improve its speed and convergence in finding the optimum solution to a well placement optimization problem. The effects of variations of the control parameters for both algorithms were studied, as well as the algorithms’ performances in the cases studied. The modified ABC (MABC) algorithm gave better results than the Artificial Bee Colony algorithm. It was noticed that the performance of the ABC algorithm increased with increase in the number of its optimization agents for both algorithms studied. The modified ABC algorithm overcame the challenge posed by the use of uniformly generated random numbers with very rough NPV surface. This new modified ABC algorithm proposed in this work will be a great tool in optimization for the Petroleum industry as it involves Well placements for optimum oil production.

A Computational Comparison between Optimization Techniques for Wells Placement Problem: Mathematical Formulations, Genetic Algorithms and Very Fast Simulated Annealing

This study considers several computational techniques for solving one formulation of the wells placement problem (WPP). Usually the wells placement problem is tackled through the combined efforts of many teams using conventional approaches, which include gathering seismic data, conducting real-time surveys, and performing production interpretations in order to define the sweet spots. This work considers one formulation of the wells placement problem in heterogeneous reservoirs with constraints on inter-well spacing. The performance of three different types of algorithms for optimizing the well placement problem is compared. These three techniques are: genetic algorithm, simulated annealing, and mixed integer programming (IP). Example case studies show that integer programming is the best approach in terms of reaching the global optimum. However, in many cases, the other approaches can often reach a close to optimal solution with much more computational efficiency.

智能反射面室内部署的位置规划方法

智能反射面(IRS/RIS)应用于实际无线通信系统时,如何优选布放位置及面板取向,是提升技术实效所面临的主要问题之一。RIS布放的数学规划问题,不仅有优化目标的设计,还要考虑通信环境的建筑物分布和可选部署面的有效散射截面。相比于室外较为开放的空间,RIS的室内部署存在更多限制性条件。针对室内既有墙面的RIS布放,该文设计了多终端接入选址规划模型和等价问题。为约化其非线性计算,通过单终端退化分析,给出目标函数的卡西尼卵形线分布特征,证明RIS布放范围限于各终端及基站在部署面投影的重叠区,提出基于对半搜索法的高效启发式算法。数值仿真了2种复杂室内结构的多终端接入场景,结果表明所提算法不仅有显著加速效益,还可扩展用于多RIS网络规划。

双目标优化与生成对抗网络结合的框架结构阻尼器布置方案智能设计方法

为实现框架结构的阻尼器智能化布置,结合减震设计原理和智能算法,采用双目标优化算法和生成对抗网络算法分别进行阻尼器竖向和水平智能布置研究,并将该方法应用到两个框架结构减震设计工程案例中。在框架结构减震设计中,采用双目标优化算法进行阻尼器竖向布置,并与逐层逼近法、工程师设计和非减震设计进行对比,结果表明,采用该优化算法得到的阻尼器竖向布置方案能有效降低层间位移角和楼层加速度,提高结构的抗震性能。在确定各楼层的阻尼器数量后,利用训练好的生成对抗网络生成模型,可快速、自动地选择和确定各楼层阻尼器的平面安装位置,生成的平面布置与工程师设计的平面布置在相似性差异度综合评价指标上小于临界值0.1,说明两者相似度较高,且有利于提高原结构的抗扭能力。将双目标优化算法与生成对抗网络相结合,不仅能满足框架结构的减震性能目标,而且可实现阻尼器布置方案的智能设计,提升减震工程设计效率。

一种联合边缘服务器部署与服务放置的方法

边缘计算(EC)在靠近用户的网络边缘部署边缘服务器(ES),并将服务放置在ES上,从而可以满足用户的服务需求。独立研究ES部署和服务放置问题的成果已有很多,但两者存在高度耦合关系。考虑到EC系统的收益,有必要提供付费服务,使得EC系统处理用户服务请求时会获得相应收入。同时,EC系统处理用户服务请求时会产生时延和能耗成本,为了最大化EC系统的收益,在用户服务请求和服务价格不同的约束下,需要合适的服务放置方案来提高EC系统的收益。为此,在ES与基站之间的位置关系、ES部署和服务放置之间的耦合关系、服务副本数和服务价格等约束下,提出一种包括改进的k-means算法和多智能体强化学习算法的两步方法,使EC系统的收益最大化。首先,构建一个联合ES部署和服务放置模型,其中ES部署明确考虑了基站之间的位置关系,服务放置明确考虑了ES部署的位置,以及不同的服务请求和价格;然后,基于基站的位置关系和基站的服务请求负载,通过带约束的k-means算法,在不同约束条件下分别确定最佳的ES部署位置以及ES的协作域;最后,以最大化EC系统收益为目标,通过多智能体强化学习算法在ES上放置服务。实验结果表明,与对比方法相比,所提方法能够提高收益7%~23%。

非规则再生骨料建模及再生骨料混凝土数值模拟

传统再生骨料混凝土(RAC)模型多采用基本几何形状来代表再生骨料(RA)结构,这与实际情况差异较大。本文结合数字图像技术和傅里叶描述法构建了真实的二维骨料数据库,并提出了基于骨料叠加法随机生成RA的新方法。基于外接矩形框/内切圆的粗判定和基于重叠框的精细判定提出了一种新的适用于傅里叶重构骨料的快速投放算法,实现了不同RA取代率和RA残余砂浆含量的RAC细观结构的构建。基于内聚力模型根据随机生成的RAC的细观结构建立了数值模型,研究了RA取代率和RA残余砂浆含量对RAC单轴抗拉性能的影响。结果表明,RAC的抗拉强度和弹性模量随着RA取代率和RA残余砂浆含量的增加而降低,但峰值应变变化不大。该RA模型更符合实际,且附着砂浆的含量和分布可自主调整。

基于分组遗传算法的数据中心虚拟机节能映射

近年来,随着人们对云计算业务需求持续增长,数据中心能耗日益增加,由此不仅增加了运营成本,巨大的碳排放对生态环境也产生严重的影响,数据中心节能已成为当前亟须解决的重要难题。对云数据中心的虚拟机放置(Virtual Machine Placement,VMP)进行优化能有效地提高资源利用率,同时,VMP也是减少数据中心能耗的重要技术之一;针对数据中心的能耗感知VMP问题,提出一种基于分组遗传算法的节能算法EEGGA(Energy-Efficient Grouping Genetic Algorithm),算法将节能VMP问题视为装箱问题(Bin Packing Problem,BPP),并应用基于分组编码的遗传算法对其进行求解,通过减少活动物理主机的数量(装箱数量)以实现降低数据中心能耗的目标;在算法迭代过程的交叉和变异等阶段,设计了多种启发优化策略提升子代染色体的适应度,从而提高算法的节能性能和加快迭代收敛的速度;通过仿真实验,在收敛速度和求解性能等方面将提出的算法与传统的节能遗传算法进行对比,实验结果表明:提出的算法能够有效地减少数据中心的能耗,在节能性能和求解收敛速度方面均优于其他算法。

基于B^(*)-树的连续微流控生物芯片流层物理设计方法

作为连续微流控生物芯片设计中的关键步骤,流层物理设计是一个复杂的多目标优化问题,传统方法通常将其分为布局和布线2个阶段。为弥补这2个阶段之间的交互不足,提出了一种基于B^(*)-树的流层物理设计方法,以提高解的质量和求解效率。在布局阶段,采用Memetic算法更高效地搜索解空间。在布线阶段,使用考虑通道顺序的基于协商的算法,同时将布线信息反馈给布局阶段,以调整布局并减少拥挤区域和通道交叉。实验结果表明,与现有算法相比,所提算法在芯片面积、通道总长度以及通道交叉数量等方面取得了更优的效果。

Large Scale VLSI Module Placement Using LFF Heuristics by Stages

We present a deterministic algorithm for large-scale VLSI module placement. Following the less flexibility first （LFF） principle,we simulate a manual packing process in which the concept of placement by stages is introduced to reduce the overall evaluation complexity. The complexity of the proposed algorithm is （N1 ＋ N2 ） × O（ n^2 ） ＋ N3× O（n^4lgn） ,where N1, N2 ,and N3 denote the number of modules in each stage, N1 ＋ N2 ＋ N3 = n, and N3〈〈 n. This complexity is much less than the original time complexity of O（n^5lgn）. Experimental results indicate that this approach is quite promising.

多天线无线充电器的安全布置研究

多天线无线充电器的安全布置问题(SPINNER)涉及给定一组无线可充电设备和一组无线充电器,每个充电器都配备了多根有向天线,需确定其功率水平和策略(包括充电器的位置和天线方向),以最大化整体充电效用,并确保目标区域内任意位置均满足电磁辐射安全约束。文中考虑了两种场景,分别是具有给定位置集的多天线无线充电器的安全布置(SPINNER-G)和具有任意点的多天线无线充电器的安全布置(SPINNER-A)。首先,采用分段常数函数近似非线性充电功率函数,并将二维区域划分为有限数量的子区域,使得无限的电磁辐射安全约束减少为有限个。然后,针对SPINNER-G,提出了一种最大覆盖集提取方法,以限制无线充电器的方向。针对SPINNER-A,构造了最大相交情况,以限制充电器的位置和方向。文中提出了两种基于线性规划的贪心方案,分别求解SPINNER-G和SPINNER-A,近似比均为1/2-ε。仿真结果表明,所提算法在充电效用方面相比3种对比算法至少提高了54.2%。

考虑储能调控优化的配电网分布式电源选址定容

稳定配电网潮流分布、明确分布式电源的接入位置和容量是含分布式电源配电网优化运行的重要问题。提出一种基于深度强化学习算法的储能调控优化模型,实现分布式电源配置与用电负荷需求关系的匹配,从而稳定高渗透率下配电网的潮流分布。以线路损耗与电压波动性为损失函数,提出基于多目标遗传算法的分布式电源选址定容决策模型。在IEEE 14节点系统进行测试,结果表明该算法能够有效选择分布式电源的最佳接入位置和容量,在保证电压幅值不产生过大波动的同时,进一步降低了整体网络的线路损耗。

高频空化冲击作用下HTPB固体推进剂的细观损伤机制

为研究高频空化冲击作用下的HTPB复合固体推进剂热力耦合行为及细观损伤机制,提出一种考虑颗粒/基体间实际界面相的三维全级配HTPB固体推进剂细观模型构筑方法,区别于在颗粒/基体界面间嵌入虚拟内聚力界面单元的传统细观模型,进一步建立起空化微射流冲击作用下的固体推进剂热力耦合细观力学模型,分析了细观尺度上的固体推进剂破碎规律、损伤行为、局部应力应变和温度分布情况。结果表明,空化微射流作用在AP颗粒上后,导致AP颗粒直接破裂,随着冲击程度的增加,AP颗粒与HTPB基体间界面层受到冲击作用而破裂;固体推进剂在空化微射流冲击后的最大应力值为34.27MPa、应变值为1.314;应力波传递过程由于受到AP颗粒、Al颗粒和界面相的阻碍作用而发生传递路径的改变;空化冲击固体推进剂过程的最大温度值由于断裂能、内能和摩擦能的逐渐累积而呈现出逐步增长的变化特征,且最大温度值为24.59℃,在距离空化微射流较远位置的固体推进剂由于传热系数较低而无明显温升。

基于相量测量单元优化配置的配电网谐波状态估计研究

随着大量电力电子设备的接入,配电网谐波问题愈发严重。谐波状态估计的准确性直接影响到后续的谐波治理效果。相量测量单元(phasor measurement unit,PMU)可以实时测量节点电压与支路电流,可借助其实现谐波状态估计。然而目前PMU价格较高,如何进行合理的优化配置保证全网谐波状态可观,同时提高谐波状态估计的准确性,是亟待解决的问题。首先构建以PMU经济配置和谐波状态估计精度最高为目标的PMU优化配置模型,并提出一种改进二进制粒子群-遗传混合算法用于求解。随后在实时仿真器中搭建IEEE14节点模型,选用均值插补法以及Vondrak滤波法进行数据处理并分析了优化所得多种PMU配置场景对谐波状态估计的影响。结果表明:所提算法从减少投资成本及降低谐波状态估计误差角度考虑,能够给出合理的PMU配置方案,有助于支撑工程决策。

液体火箭发动机涡轮泵多维度传感器优化布置

针对液体火箭发动机涡轮泵健康监测时存在的结构重要模态信息遗漏、故障敏感信息提取不全等问题,提出了一种涡轮泵传感器多维度优化布置方法,并采用故障模拟实验台进行了实验验证。首先,建立了涡轮泵待优化区域的有限元模型并开展了结构约束模态分析以及轴承故障仿真瞬态动力学分析;其次,基于离散人工蜂群算法及瞬态分析结果得出传感器布置候选点集;再次,综合传感器多维度评估方法得出最终的传感器布置方法;最后,通过实验对比分析了传感器布置方法与传统方法的综合指标。计算结果表明:相对于有效独立法,布置测点优化后,传感器监测信号的模态振型符合率提高了20.1%,对轴承故障检测的准确率提高了27.5%,验证了多维度优化布置方法具有良好的故障诊断综合性能。

基于探路者算法的VNF放置与调度联合优化

虚拟网络功能放置和虚拟网络功能调度是近年来研究的热点,二者紧密关联缺一不可。将两个环节统一考虑、全局寻优既是技术发展趋势,也是网络功能虚拟化落地面临的艰巨挑战之一。然而,现有研究多将二者分立考虑,仅着眼单一问题的研究缺乏整体观。论文提出了一种动态的放置调度联合优化模型,在满足相关约束下,最小化服务功能链的完成时间。提出了一种改进的探路者算法解决该问题,引入了群体划分的机制增强全局搜索能力避免陷入局部最优。进一步设计了个体交叉和变异机制,减小对个体解的破坏,增强了局部搜索能力以及种群的多样性。此外,精英保留策略加快了算法的收敛。实验数据表明,论文提出的算法相较于各种主流进化算法在服务功能链完成时间评价指标上均有显著优势。