云计算中虚拟机资源与应用系统参数的协同自适应配置研究

项目名称： 云计算中虚拟机资源与应用系统参数的协同自适应配置研究

项目编号： No.61272382

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 自动化技术、计算机技术

项目作者： 彭志平

作者单位： 广东石油化工学院

项目金额： 80万元

中文摘要： 瞬息万变的云环境客观上要求虚拟机资源和应用系统参数以协同方式进行在线、高效自适应配置，以确保在满足云用户的服务级别协定前提下，最大限度地提高云服务提供商的资源效用。本课题研究两者协同自适应配置的理论问题、关键技术和方法，重点研究：(1)研究以虚拟机簇为粒度的多agent组织，包括组织结构和工作流。(2)研究分布式协同agent的语义一致性、非单调性，包括体系结构及其语义模型、描述逻辑与可废止逻辑理论集成的协同配置知识库、不一致本体转换为可废除逻辑规则的高效映射算法、可废止逻辑理论推理规则的扩展、知识冲突处理机制以及非单调的知识表征与推理。(3)研究分布式自适应学习agent的高性能学习机制，包括基于强化学习理论与方法的动作协调与约简、状态抽象和知识迁移，多途径提高其自适应学习性能。(4)研制原型系统，构建试验床，验证研究成果的可适用性和可扩展性。

中文关键词： 云计算；协同自适应配置；虚拟机；应用系统；

英文摘要： In the ever-changing cloud computing environments, it is objectively required that both the virtual machine resources and the hosted application system parameters should adaptively be configured online and efficiently in a collaborative manner, in order to maximize cloud service providers' resource utilization when cloud users' service level agreement is met. This project investgates the theoretical issues, key technologies and methods of their collaborative adaptive configurations, focusing on: (1)investigating the multi-agent organization in a granularity of a virtual machine cluster, including its organizational structure and workflow.(2)investigating the semantic consistency and non-monotonic of the distributed collaborative agent, including its architecture and semantic model, the collaborative configuration knowledge bases based on integrating description logic with defeasible logic theory, the highly efficient mapping algorithm for inconsistent ontology transforming to the defeasible logic rules, the extension of inference rules of the defeasible logic theory, the conflict resolution mechanisms of knowledge, and the representation and reasoning of non-monotonic knowledge.(3)investgating the high-performance learning mechanisms of the distributed adaptive learning agent, including action set coordination a

英文关键词： Cloud Computing；Collaborative Adaptive Configuration；Virtual Machine；Application System；