Scheduling Policies for Processor Coallocation in Multicluster Systems

Building multicluster systems out of multiple, geographically distributed clusters interconnected by high-speed wide-area networks can provide access to a larger computational power and to a wider range of resources. Jobs running on multiclusters and, more generally, in grids, may require (processor) coallocation, i.e., the simultaneous allocation of resources (processors) in different clusters or subsystems of a grid. In this paper, we propose four scheduling policies for processor coallocation in multiclusters, and we assess with simulations their performance under a wide variety of parameter settings. In particular, in our simulations we use synthetic workloads and workloads derived from the logs of actual systems and from runtime measurements. We conclude that although coallocation makes scheduling more difficult and the wide-area communication critically impacts the performance, there is a wide range of realistic applications that may benefit from coallocation. However, unrestricted coallocation is not recommended: Limiting the total job size or the number or the sizes of their components improves performance.     

Stable Model Predictive Control for Constrained Max-Plus-Linear Systems

Discrete-event systems with synchronization but no concurrency can be described by models that are “linear” in the max-plus algebra, and they are called max-plus-linear (MPL) systems. Examples of MPL systems often arise in the context of manufacturing systems, telecommunication networks, railway networks, parallel computing, etc. In this paper we provide a solution to a finite-horizon model predictive control (MPC) problem for MPL systems where it is required that the closed-loop input and state sequence satisfy a given set of linear inequality constraints. Although the controlled system is nonlinear, by employing results from max-plus theory, we give sufficient conditions such that the optimization problem that is performed at each step is a linear program and such that the MPC controller guarantees a priori stability and satisfaction of the constraints. We also show how one can use the results in this paper to compute a time-optimal controller for linearly constrained MPL systems.

不确定系统的稳定广义预测控制

针对一类有界不确定线性离散被控对象,采用Min-Max优化方法,提出一种新的稳定广义预测控制(MMSGPC)算法.首先引入内模控制结构,将干扰和不确定性从被控对象中分离出来,并利用局部反环节对其进行补偿,然后采用Min-Max优化方法,将终端约束条件转化为有界不确定性最差情况对应的线性方程,最后通过引入矩阵的Moore-Penrose逆,得到了终端约束线性方程的通解,并结合性能指标函数求得了最优控制律.仿真实例验证了该方法的稳定效果.     

Scalable Variational Integrators for Constrained Mechanical Systems in Generalized Coordinates

We present a technique to implement scalable variational integrators for generic mechanical systems in generalized coordinates. Systems are represented by a tree-based structure that provides efficient means to algorithmically calculate values (position, velocities, and derivatives) needed for variational integration without the need to resort to explicit equations of motion. The variational integrator handles closed kinematic chains, holonomic constraints, dissipation, and external forcing without modification. To avoid the full equations of motion, this method uses recursive equations, and caches calculated values, to scale to large systems by the use of generalized coordinates. An example of a closed-kinematic-chain system is included along with a comparison with the open-dynamics engine (ODE) to illustrate the scalability and desirable energetic properties of the technique. A second example demonstrates an application to an actuated mechanical system.      

不确定系统的稳定广义预测控制

针对一类有界不确定线性离散被控对象,采用Min-Max优化方法,提出一种新的稳定广义预测控制(MMSGPC)算法.引入内模控制结构,将干扰和不确定性从被控对象中分离出来,并利用局部反环节对其进行补偿;采用Min-Max优化方法,将终端约束条件转化为有界不确定性最差情况时应的线性方程;通过引入矩阵的Moore-Penrose逆,得到了终端约束线性方程的通解,并结合性能指标函数求得了最优控制律.通过仿真实例验证了该方法的稳定效果.     

多处理器混合关键性系统中的划分调度策略

多核处理器正越发广泛地应用到现代嵌入式系统的设计与实现当中,其强大的计算能力为将多个不同关键性级别的功能子系统集成到统一的共享资源平台提供了支持.混合关键性系统的调度问题即便在单处理器平台中都极具挑战性,在多处理器平台则更为困难.将目前资源利用率最高的单处理器混合关键性调度算法EY-VD扩展到多处理器平台中.首先,结合传统的划分调度策略提出了适用于多处理器混合关键性系统的MC-PEDF(mixedcriticality partitioned earliest deadline first)划分调度算法.尽管比之前的算法有更好的可调度性能,但传统的划分策略不能有效地平衡不同关键性级别下的负载,故其不完全适用于混合关键性系统.为了克服传统策略的不足,提出了划分调度策略OCOP(one criticality one partition).OCOP允许系统在关键性模式切换时对实时任务集进行重新划分,进而更好地平衡各个处理器在不同关键性模式中的资源利用率.基于OCOP,提出了第2种划分调度算法MC-MP-EDF(mixed-criticality multi-partitioned EDF).基于随机生成任务集的仿真实验结果表明,与MC-PEDF和已有的算法相比,MC-MP-EDF能够显著地提高系统的可调度性,尤其是在处理器数量较多的系统中.     

Generalized Algebraic Deadlock Avoidance Policies for Sequential Resource Allocation Systems

Currently, one of the most actively researched approaches regarding the design of deadlock avoidance policies for sequential resource allocation systems is based on concepts and techniques provided by the, so called, theory of regions, that addresses the broader problem of synthesizing PN models with prespecified behaviors. However, one limitation of the theory of regions and its aforementioned derivatives is that they cannot be applied when the target behavior has a nonconvex representation in the underlying state space. In this note, we show how this problem can be circumvented by appropriately generalizing the employed class of the candidate policies.     

Performance Analysis of Retrial Queueing Systems Using Generalized Stochastic Petri Nets

We consider retrial queueing systems, in which an arriving customer finding the server busy, may repeat his call after a random duration. The consideration of repeated calls introduces great analytical difficulties. In fact, detailed analytical results exist for some special retrial queueing systems, while for many others, the performance evaluation is limited to numerical algorithms, approximation methods and simulation. Retrial queues have been widely used to model many problems in telephone switching systems, telecommunication and computer networks. In this paper, we show a method of modelling and analysing retrial queueing systems, using the Generalized Stochastic Petri nets (GSPNs).     

一类广义不确定线性系统稳定控制

本文讨论一类广义不确定线性系统Ex’(t)=(A+δA)x(t)+(B+δB)u(t)的稳定控制问题,这里，E为奇异阵；δA∈UA(α) δB∈UB(β)均为系统的不稳定量，UA(α)、UB(β)为某类集合，α>0, 0<β<（√2-1）。在广义线性系统Ex’(t)=Ax(t)+Bu(t)能稳、能正常化条件下，上述不确定系统具有稳定鲁棒控制器，并且该控制器可消除该系统的脉冲行为。     

Monotonicity Properties for Multi-Class Queueing Systems

We study multi-dimensional stochastic processes that arise in queueing models used in the performance evaluation of wired and wireless networks. The evolution of the stochastic process is determined by the scheduling policy used in the associated queueing network. For general arrival and service processes, we give sufficient conditions in order to compare sample-path wise the workload and the number of users under different policies. This allows us to evaluate the performance of the system under various policies in terms of stability, the mean overall delay and the mean holding cost. We apply the general framework to linear networks, where users of one class require service from several shared resources simultaneously. For the important family of weighted α-fair policies, stability results are derived and monotonicity of the mean holding cost with respect to the fairness parameter α and the relative weights is established. In order to broaden the comparison results, we investigate a heavy-traffic regime and perform numerical experiments. In addition, we study a single-server queue with two user classes, and show that under Discriminatory Processor Sharing (DPS) or Generalized Processor Sharing (GPS) the mean overall sojourn time is monotone with respect to the ratio of the weights. Finally we extend the framework to obtain comparison results that cover the single-server queue with an arbitrary number of classes as well.     

Approximation Algorithms for Time Constrained Scheduling

In this paper we consider the following time constrained scheduling problem. Given a set of jobsJwith execution timese(j)(0, 1] and an undirected graphG=(J, E), we consider the problem to find a schedule for the jobs such that adjacent jobs (j, j′)Eare assigned to different machines and that the total execution time for each machine is at most 1. The goal is to find a minimum number of machines to execute all jobs under this time constraint. This scheduling problem is a natural generalization of the classical bin-packing problem. We propose and analyse several approximation algorithms with constant absolute worst case ratio for graphs that can be colored in polynomial time.     

求解广义优先关系下多技能人员项目调度问题的改进布谷鸟搜索算法

为解决传统的完成-开始时序不能满足描述真实项目调度顺序要求的问题,引入广义优先关系(GPRs)及改进的AON描述任务的时序约束。提出将布谷鸟搜索算法应用于求解广义优先关系下的多技能人力资源项目调度问题(MS-RCPSP/GPRs)中的构想,建立了基于改进布谷鸟搜索算法(ICS)的求解方法,采用Powell局部改进技术和精英保留策略,并给出了算法流程。基于相关案例生成器生成该问题的数据集,实验结果表明ICS是一种求解MS-RCPSP/GPRs的有效方法,对解决实际问题具有重要意义。     

即时调度中周期调度最大化的带宽预留策略

越来越多的高性能网络通过电路交换或MPLS/GMPLS技术提供专用信道,支持大数据传输。为带宽预留服务开发有效的调度算法已成为提高网络资源利用率和满足应用用户传输要求的关键任务。高性能网络中即时带宽的研究集中关注在单次性能,本文对于即时调度中的周期性能优化,考虑一个新的问题:即时调度中的周期调度最大化问题。本文证明此问题是NP问题,针对此问题提出并实现了一个启发式算法:FBMHA,对FBMHA与Greed-MSR算法进行了大量的实验进行评估。实验结果表明,FBMHA算法相比于Greed-MSR算法在成功率和传输数据量方面有大的提升,表现出了FBMHA算法的优越性。     

Perturbation Analysis for Stochastic Fluid Queueing Systems

Recent study for congestion control in high speed networks indicates that the derivative information for the congestion at the common buffer for multiple sources could be useful in achieving efficient and fair allocation of the bandwidth (Kelly, 1997; Kelly et al., 1998). In this paper we present an algorithm that estimates such derivatives for multiple on-off sources. The algorithm has its root in the infinitesimal perturbation analysis (IPA) for the classical queueing systems. Although the traditional IPA algorithm does not give unbiased derivative estimates for multi-class arrivals, we are able to prove the unbiasedness in the case of multi-class ON-OFF sources. The results in this paper may motivate a new look at the end-to-end congestion control issue.     

能量捕获传感网中吞吐量最大化的占空比方案

对于以超级电容为储能设备且工作在占空比模式下的能量捕获传感网,其节点的平均能量捕获速率和吞吐量均很大程度上取决于节点开始休眠时的电容电压和醒来工作时的电容电压。首先,通过理论分析证明传感器节点的能量捕获速率很大程度上取决于电容的即时电压值;然后,对节点的最大吞吐量问题进行建模;最后,提出一种最优化方案来确定开始休眠时的电容电压和醒来工作时的电容电压的最佳取值,从而最大化节点的吞吐量。实验表明,所提方案的吞吐量比已有占空比方案的吞吐量有明显的提高。     

Effective Routing and Scheduling in Adversarial Queueing Networks

In an adversarial queueing network, the incoming traffic is decided by an adversary, who operates under a reasonable rate restriction. This model provides a valuable, complementary point of view to that of the traditional queueing network model in which arrivals are modeled by stochastic processes. As a result, the adversarial queueing network model has attracted much attention in recent years, especially as a way of modeling packet injections into a communication network. Our main result is a simple, effective packet routing and scheduling algorithm with a provably good performance. Specifically, our algorithm keeps the system stable (bounded number of packets in the system), with the bound on the number of packets in the system that is O((1 - r)^-1), where r can be interpreted as the arrival rate of the packets into the communication network. This improves upon the work of Gamarnik, who designed an algorithm for which the number of packets in the system is O((1 - r)^-2); moreover, our result matches the traditional queueing-theoretic number-in-system bound.     

基于状态空间的有约束稳定广义预测控制

针对一类有约束的稳定广义预测控制问题,提出一种基于状态空间的稳定广义预测控制算法。首先通过传递函数的状态空间实现,得到被控对象的离散状态空间形式;然后引入Deadbeat状态反馈矩阵并给出约束条件的等价性定理,实现了约束条件的等价转化;最后通过等价约束条件优化性能指标函数求解控制律。仿真实例表明该方法具有良好的稳定性。     

吞吐量和延时约束下的片上通信结构的Pareto空间优化

SoC中各IP核之间的互连结构是决定片上系统性能的关键因素.近年来,片上互连通信结构的配置与优化成为SoC通信综合的研究重点和热点,而已有方法优化SoC互连通信结构的仿真速度较慢,支持设计自动化的能力较差,使用的单目标优化算法无法解决多个性能目标之间的冲突.针对以上不足提出了吞吐量和延时约束下的片上互连通信结构的自动配置与优化的方法,该方法提出了片上总线互连通信结构模板,使用事务级通信仿真和多目标演化算法,探索吞吐量和延时约束下的多目标Pareto空间.与已有的先进Srinivasan方法相比,该方法的吞吐量提高10%,传输延迟降低17%,有效提高了SoC互连通信结构的优化质量.     

Stable and Real-Time Scheduling of a Class of Hybrid Dynamic Systems

Based on some practical engineering problems arising from process control and space-structure control, this paper studies a class of hybrid dynamic systems in which N plants are controlled by a central controller in sharing time manner, where the plants are described by differential equations and the controller works according to the mechanism of discrete events. An event feedback strategy is suggested to be a scheduling policy such that one and only one plant among N plants is chosen to be controlled at any time. Some conditions of asymptotical and exponential stability are then given and an exponential upper bound of states norm is also estimated for the event feedback scheduling strategy. An algorithm based on event feedback strategy is presented to determine the control laws of the plants to meet the given performance. An example follows to illustrate the application and effect of the results.