基于异步奖励深度确定性策略梯度的边缘计算多任务资源联合优化

移动边缘计算（MEC）系统中,因本地计算能力和电池能量不足,终端设备可以决定是否将延迟敏感性任务卸载到边缘节点中执行。针对卸载过程中用户任务随机产生且系统资源动态变化问题,提出了一种基于异步奖励的深度确定性策略梯度（asynchronous reward deep deterministic policy gradient,ARDDPG）算法。不同于传统独立任务资源分配采用顺序等待执行的策略,该算法在任务产生的时隙即可执行资源分配,不必等待上一个任务执行完毕,以异步模式获取任务计算奖励。ARDDPG算法在时延约束下联合优化了任务卸载决策、动态带宽分配和计算资源分配,并通过深度确定性策略梯度训练神经网络来探索最佳优化性能。仿真结果表明,与随机策略、基线策略和DQN算法相比,ARDDPG算法在不同时延约束和任务生成率下有效降低了任务丢弃率和系统的时延和能耗。     

基于任务特征的云计算资源分配策略

云计算中存在大量的服务资源,高效利用这些资源是资源分配的主要任务。通过对任务流的资源占用和执行过程建模分析,掌握任务流特性,提出伙伴关系任务流资源分配方案。多任务流执行过程中出现资源需求增减的时间同步,这种现象持续不断将产生波动共振,本文采用资源申请-仲裁机制来解决这一问题,把共振的任务流转变为伙伴关系任务流来协作使用资源,保持资源的高效利用。实验结果表明,本文算法消耗的资源比静态方式要少,溢出数也要少,实现了资源的充分利用,验证了算法的正确性。     

基于概率模型检验的云渲染任务调度定量验证

云渲染技术已被广泛应用于影视和动漫等行业.与传统的渲染农场和租赁市场模式不同，云渲染系统依托云计算基础设施提供多种软件服务进行渲染作业的方式，正逐渐成为新兴的计算模式.由于任务执行和资源操作等作业调度对于用户而言是透明的，这要求云渲染系统应具备智能化以实现计算资源优化调度和多端任务管理，并对系统可靠性提出了更高要求.针对这一问题，提出了采用概率模型检验对云渲染系统任务调度进行定量评估.首先，考虑渲染服务失效等因素引发的随机系统异常和指令错误，如文件损坏和渲染任务超时等，提出了基于离散马尔可夫链（DTMC）的概率模型对云渲染系统的文件准备模块、资源请求模块、渲染任务执行模块进行形式化建模；其次，从服务质量属性角度提出了9类验证性质用于定义云渲染系统的可靠性，采用概率计算树逻辑（PCTL）描述检验性质公式并执行工具PRISM计算和验证渲染系统可靠性；最后，结合案例和实验证明了该方法的可行性和有效性，尤其是对改进前后云渲染系统进行定量检验，可用于指导如何进行失效恢复和任务切换.因此，该方法在一定程度上可提高云渲染系统的可靠性.     

基于优先级时间Petri网的实时嵌入式多核系统分析

已有的基于点区间优先级时间Petri网分析实时嵌入式多核系统的工作, 存在以下不足: (1)点区间优先级时间Petri网只考虑每个任务的执行时间是一个固定值的情况, 而更多的实际应用中每个任务的执行时间是在一个区间范围内, 因此不能模拟这些应用; (2)没有实现从任务依赖图到点区间优先级时间Petri网的自动转化, 不便于工程设计人员使用; (3)没有考虑任务间互斥访问共享变量的情况. 为此, 定义了优先级时间Petri网(Pri-TPN)以弥补第1个不足; 定义带有资源分配与优先级的任务依赖图(TDG-RAP)以弥补第3个不足; 给出从TDG-RAP到Pri-TPN的转化规则与算法以弥补第2个不足, 以及基于Pri-TPN分析任务最坏执行时间与系统死锁的算法; 开发工具软件, 方便工程设计人员使用.     

Online optimization for scheduling preemptable tasks on IaaS cloud systems

In Infrastructure-as-a-Service (IaaS) cloud computing, computational resources are provided to remote users in the form of leases. For a cloud user, he/she can request multiple cloud services simultaneously. In this case, parallel processing in the cloud system can improve the performance. When applying parallel processing in cloud computing, it is necessary to implement a mechanism to allocate resource and schedule the execution order of tasks. Furthermore, a resource optimization mechanism with preemptable task execution can increase the utilization of clouds. In this paper, we propose two online dynamic resource allocation algorithms for the IaaS cloud system with preemptable tasks. Our algorithms adjust the resource allocation dynamically based on the updated information of the actual task executions. And the experimental results show that our algorithms can significantly improve the performance in the situation where resource contention is fierce.     

A task‐based and resource‐aware approach to dynamically generate optimal software architecture for intelligent service robots

Intelligent service robots provide various services to users by understanding the context and goals of a user task. In order to provide more reliable services, intelligent service robots need to consider various factors, such as their surrounding environments, users' changing needs, and constrained resources. To handle these factors, most of the intelligent service robots are controlled by a task‐based control system, which generates a task plan that represents a sequence of actions, and executes those actions by invoking the corresponding functions. However, the traditional task‐based control systems lack the consideration of resource factors even though intelligent service robots have limited resources (limited computational power, memory space, and network bandwidth). Moreover, system‐specific concerns such as the relationships among functional modules are not considered during the task generation phase. Without considering both the resource conditions and interdependencies among software modules as a whole, it will be difficult to efficiently manage the functionalities that are essential to provide core services to users. In this paper, we propose a mechanism for intelligent service robots to efficiently use their resources on‐demand by separating system‐specific information from task generation. We have defined a sub‐architecture that corresponds to each action of a task plan, and provides a way of using the limited resources by minimizing redundant software components and maintaining essential components for the current action. To support the optimization of resource consumption, we have developed a two‐phase optimization process, which is composed of the topological and temporal optimization steps. We have conducted an experiment with these mechanisms for an infotainment robot, and simulated the optimization process. Results show that our approach contributed to increase the utilization rate by 20% of the robot resources. Copyright © 2011 John Wiley & Sons, Ltd.     

Reinforcement learning based resource allocation in business process management

Efficient resource allocation is a complex and dynamic task in business process management. Although a wide variety of mechanisms are emerging to support resource allocation in business process execution, these approaches do not consider performance optimization. This paper introduces a mechanism in which the resource allocation optimization problem is modeled as Markov decision processes and solved using reinforcement learning. The proposed mechanism observes its environment to learn appropriate policies which optimize resource allocation in business process execution. The experimental results indicate that the proposed approach outperforms well known heuristic or hand-coded strategies, and may improve the current state of business process management.     

基于边缘计算的新型任务卸载与资源分配策略

针对移动边缘计算（MEC）中密集型任务卸载时，系统开销较大和延时抖动明显的问题，提出一种新型资源分配策略。首先在系统时延约束下，分析了系统任务执行开销与终端设备的资源分配机制；其次建立了基于计算卸载和任务分配的联合凸优化目标；最后采用拉格朗日乘子法进行迭代更新得到最优解。仿真结果表明，所提任务卸载与资源分配方案在保证用户服务质量的同时降低了任务执行开销，并有效提升了MEC系统性能。     

移动边缘计算中基于Lyapunov的任务卸载与资源分配算法

移动边缘计算（MEC）通过将计算和存储资源部署在无线网络边缘,使得用户终端可将计算任务卸载到边缘服务器进行处理,从而缓解终端设备资源受限与高性能任务处理需求之间的冲突。但随着任务卸载规模的不断增加,执行任务所产生的功耗急剧上升,严重影响了MEC系统的收益。建立任务队列动态调度模型,以队列上溢概率为约束构建最大化系统平均收益的资源优化模型。考虑到资源优化问题为不同时隙下的耦合问题,运用Lyapunov优化理论设计一种基于单时隙的资源分配算法,将优化问题转化为用户本地计算资源分配、功率和带宽资源分配以及MEC服务器计算资源分配3个子问题并分别进行求解。仿真结果表明,该算法在满足用户QoS需求的同时能够有效提高MEC系统的时间平均收益。     

Contextual Resource Negotiation-Based Task Allocation and Load Balancing in Complex Software Systems

In the complex software systems, software agents always need to negotiate with other agents within their physical and social contexts when they execute tasks. Obviously, the capacity of a software agent to execute tasks is determined by not only itself but also its contextual agents; thus, the number of tasks allocated on an agent should be directly proportional to its self-owned resources as well as its contextual agents' resources. This paper presents a novel task allocation model based on the contextual resource negotiation. In the presented task allocation model, while a task comes to the software system, it is first assigned to a principal agent that has high contextual enrichment factor for the required resources; then, the principal agent will negotiate with its contextual agents to execute the assigned task. However, while multiple tasks come to the software system, it is necessary to make load balancing to avoid overconvergence of tasks at certain agents that are rich of contextual resources. Thus, this paper also presents a novel load balancing method: if there are overlarge number of tasks queued for a certain agent, the capacities of both the agent itself and its contextual agents to accept new tasks will be reduced. Therefore, in this paper, the task allocation and load balancing are implemented according to the contextual resource distribution of agents, which can be well suited for the characteristics of complex software systems; and the presented model can reduce more communication costs between allocated agents than the previous methods based on self-owned resource distribution of agents.     

The analytic hierarchy process: task scheduling and resource allocation in cloud computing environment

Resource allocation is a complicated task in cloud computing environment because there are many alternative computers with varying capacities. The goal of this paper is to propose a model for task-oriented resource allocation in a cloud computing environment. Resource allocation task is ranked by the pairwise comparison matrix technique and the Analytic Hierarchy Process giving the available resources and user preferences. The computing resources can be allocated according to the rank of tasks. Furthermore, an induced bias matrix is further used to identify the inconsistent elements and improve the consistency ratio when conflicting weights in various tasks are assigned. Two illustrative examples are introduced to validate the proposed method.     

基于QMix的车辆云计算资源动态分配方法

城市交通智能化和通信技术的进步会产生大量基于车辆的应用,但目前车辆有限的计算资源无法满足车辆应用的计算需求与延迟性约束。车辆云（VC）可以高效地调度资源,从而显著降低任务请求的延迟与传输成本。针对VC环境下任务卸载与计算资源分配问题,提出一个考虑异质车辆和异质任务的计计资源分配算法。对到达的任务构建M/M/1队列模型与计算模型,并定义一个效用函数以最大化系统整体效用。针对环境中车辆地理分布的高度动态系统变化,提出基于双时间尺度的二次资源分配机制（SRA）,使用两个不同时间尺度的资源分配决策动作,对其分别构建部分可观测马尔可夫决策过程。两个决策动作通过执行各自的策略获得的奖励进行连接,将问题建模为两层计算资源分配问题。在此基础上提出基于二次资源分配机制的多智能体算法SRA-QMix求解最优策略。仿真结果表明,与深度确定性策略梯度算法对比,该算法的整体效用值和任务完成率分别提高了70%、6%,对于QMix和MADDPG算法分别应用SRA后的任务完成率分别提高了13%与15%,可适用于动态的计算资源分配环境。     

一种云环境中的动态细粒度资源调度方法

云计算平台中普遍采用固定资源量的粗粒度资源分配方式,由此会引起资源碎片、过度分配、低集群资源利用率等问题.针对此问题,提出一种细粒度资源调度方法,该方法根据相似任务运行时信息推测任务资源需求;将任务划分为若干执行阶段,分阶段匹配资源,从分配时间和分配资源量两方面细化资源分配粒度;资源匹配过程中,基于资源可压缩特性进一步提高资源利用率和性能;采用资源监控、策略调整、约束检查等机制保证资源使用效率和负载性能.在开源云资源管理平台中,基于细粒度资源调度方法实现了调度器.实验结果表明：细粒度资源调度方法可以在不丧失公平性且调度响应时间可接受的前提下,细化资源匹配的粒度,有效提高云计算平台资源利用率和性能.     

The Process Enactment Tool Framework—Transformation of software process models to prepare enactment

Rich development process models contain information about structures for project organization and also for concrete outcomes of a project. However, rich processes are hard to implement. They often contain hundreds of pages of documentation. Development teams tend to be skeptical about rich processes in fear of additional effort, risking the benefits of rich tool support for enactment. Process enactment is a challenging task. There is no common methodology to quickly “implement” a development process in a tool or a set of tools. Often specialized tools are used to provide assistance during the project and it is the project manager’s task to consolidate the information with the rest of the team.The Process Enactment Tool Framework (PET) is a software tool that supports the transformation of a given formal development process into a format that project tools can work with. PET is an instrument to import processes based on a metamodel and provide exports for a specific project environment. PET takes an input software development process model and transforms it into an intermediate format that serves as the basis for a second transformation step into data formats of tools such as office suites or comprehensive ALM platforms. In this paper we present the tool framework and show how metamodel-based processes can be transformed into an environment that is ready to use for a project team. We show how PET is applied for the German V-Modell XT and for SPEM-based processes to generate, e.g., process templates for the Team Foundation Server or work product document templates.     

Developing security-aware resource management strategies for workflows

This paper investigates the resource allocation problem for a type of workflow in pervasive computing. These workflows are abstracted from the enterprise-level applications in the business or commerce area. The activities in these workflows require not only computing resources, but also human resources. Human involvement introduces additional security concerns. When we plan/allocate resource capacities, we often assume that when a task is allocated to a resource, the resource will accept the task and start the execution once the processor becomes available. However, the security policies impose further constraints on task executions, and therefore may affect both application- and system-oriented performance. Authorization is an important aspect in security. This paper investigates the issue of allocating resources for running workflows under the role-based authorization control, which is one of the most popular authorization mechanisms. By taking into account the authorization constraints, the resource allocation strategies are developed in this paper for both human resources and computing resources. In the allocation strategy for human resources, the optimization equation is constructed subject to the constraint of the budget available to hire human resources. Then the optimization equation is solved to obtain the number of human resources allocated to each authorization role. The allocation strategy for computing resources calculates not only the number of computing resources, but also the proportion of processing capacity in each resource allocated to serve the tasks assuming each role. The simulation experiments have been conducted to verify the effectiveness of the developed allocation strategies. The experimental results show that the allocation strategy developed in this paper outperforms the traditional allocation strategies, which do not consider authorization constraints, in terms of both average response time and resource utilization.     

基于效益博弈的云计算资源动态可协调分配策略研究

在对用户的任务进行计算资源分配时,为了有效提高计算资源的利用效率,减少任务执行所需要的成本,提出了一种基于效益博弈的云计算资源动态可协调分配机制。该机制采用时间矩阵和费用矩阵作为任务效益的衡量指标,提出效益博弈模型,通过该模型的效益计算方程来得到最好的资源分配策略。为了使得计算资源能够合理地按需进行分配,提出了动态可协调分配机制,在合理地分配资源,满足所有任务正常执行时所需资源的同时,最大化任务的执行效益。实验仿真及对比结果表明,在任务完成时间、任务执行的平均成本、任务完成成功率上,本文算法都取得了较好的效果。     

面向服务的无线传感器网络任务描述方式

无线传感器网络的任务协同主要指任务的描述、分解、分配、调度与执行,其中,任务描述是任务协同的基础。该文比较几种任务描述方式,针对3层的传感器网络体系结构提出一种面向服务的任务描述方式。实践表明,该描述方式提供一种简单、高效的描述手段,方便用户提交任务请求,有效地降低了任务分配系统的复杂性。     

Using clouds to scale grid resources: An economic model

Infrastructure as a Service clouds are a flexible and fast way to obtain (virtual) resources as demand varies. Grids, on the other hand, are middleware platforms able to combine resources from different administrative domains for task execution. Clouds can be used by grids as providers of devices such as virtual machines, so they only use the resources they need. But this requires grids to be able to decide when to allocate and release those resources. Here we introduce and analyze by simulations an economic mechanism (a) to set resource prices and (b) resolve when to scale resources depending on the users’ demand. This system has a strong emphasis on fairness, so no user hinders the execution of other users’ tasks by getting too many resources.Our simulator is based on the well-known GridSim software for grid simulation, which we expand to simulate infrastructure clouds. The results show how the proposed system can successfully adapt the amount of allocated resources to the demand, while at the same time ensuring that resources are fairly shared among users.     

Authoring templates with Tracker

Tracker, a tool developed by DARPA'S active templates program, extends data-entry forms to support plan authoring, information monitoring, and execution planning. Tracker is used to develop templates for a military air transportation system tool. The Tracker software lets users define templates to describe and implement a given planning problem's associated workflow.