【作者】 王倩；

【导师】 王洪峰；

【作者基本信息】 东北大学 ， 系统工程， 2019， 硕士

【摘要】 云计算是一种基于互联网共享的,为买家提供动态可扩展虚拟资源的服务模式,随着云计算的不断发展,各大企业都开始提供云计算服务,但是目前IT企业大多采用固定价格的静态定价模式,无法满足用户日益增长的多元需求,影响云提供商的收益。近年来,企业界和学术界开始探讨利用动态定价方式来解决云资源定价这一难题,其中基于组合拍卖理论的定价机制被众多学者广泛关注。由于云资源市场供需关系动态变化,导致虚拟机部署过程时变、部署成本时变,而传统的组合拍卖定价机制可能会因为仅按照竞拍价格解决胜标者问题造成云提供商的收益无法预估。为此,本文采纳系统工程的思想,面向云资源定价问题,通过分析传统的组合拍卖机制所面临的难点问题,借鉴运筹学、管理科学以及智能优化等领域的最新成果,提出考虑虚拟机部署成本的组合拍卖定价机制,对这种新型定价机制应用中的关键科学问题进行深入研究,展开了如下三个方面的研究工作:首先,相关研究工作介绍。本文从云计算、拍卖理论和云资源定价问题三个方面对相关研究进行了文献综述和成果归纳,在分析云资源市场供需关系动态变化的基础上,提出考虑虚拟机部署成本的组合拍卖定价机制的思路。其次,通过上述的分析工作,针对虚拟机部署优化问题研究。建立该问题的以部署成本最小化为目标的数学模型,并针对问题特点设计启发式求解算法,通过仿真实验验证了所提出的模型和算法的有效性。最后,在上述虚拟机部署问题的研究基础上,针对云资源定价问题研究。通过所提出的组合拍卖资源定价机制的思想,将云资源定价问题转化为考虑部署成本的胜标者确定问题,建立该问题的以收益最大化为目标的数学模型,并针对问题特点设计了基于变邻域搜索算法的求解算法,通过仿真实验验证了所提出的模型和算法的有效性。通过上述研究发现,所建立的云资源定价机制一方面能够快速有效给出云资源的价格,另一方面可以为国内IT企业的云资源定价提供一定的借鉴作用。更多还原

【Abstract】 Cloud computing is a service model based on Internet sharing that provides buyers with dynamically scalable virtualized resources.With the continuous development of cloud computing,all enterprises are beginning to provide cloud computing services.But most of the IT companies adopt a fixed-price static pricing model,which cannot meet the growing diversified demand of buyers,and affect revenues of cloud providers.In recent years,industry and academia have begun to use dynamic pricing to solve the problem of cloud resource pricing.The pricing mechanism based on the combinatorial auction theory has been widely concerned by many scholars.Due to the dynamic change of the supply and demand in the cloud resource market,the virtual machine placement process changes and placement cost changes.However,the traditional combinatorial auction pricing mechanism may cause the cloud provider’s revenue to be unpredictable because the winner determination problem is solved according to the auction price.Therefore,this thesis will study cloud resource pricing problems based on combinatorial auction in the fields of operations research,management science,intelligent computation based on the mechanism of systems engineering.The detailed research works can be summarized as follows.Firstly,this thesis presented a detailed review on the relevant research works in cloud computing,auction theory and cloud resource pricing.Based on the analysis of the dynamic changes of the supply and demand relationship in the cloud resource market,the idea of a combinatorial auction pricing mechanism considering the cost of virtual machine placement was proposed.Secondly,through the above analysis work,research on virtual machine placement problems.We established a mathematical model with the goal of minimizing placement cost,designed algorithm for the problem characteristics,and validated the effectiveness of the proposed models and algorithms by the relevant simulation experiments.Finally,combined with the research results of virtual machine placement problems,it is oriented to the study of cloud resource pricing problems.Through the idea of combinatorial auction mechanism,the cloud resource pricing problem is transformed into the winner determination problem considering the cost of placement.We established a mathematical model with the goal of maximizing revenue,and designed algorithm based on the variable neighborhood search algorithm for the problem characteristics,and validated the effectiveness of the proposed models and algorithms by the relevant simulation experiments.Based on these research works,it can be observed that the cloud resource pricing mechanism can quickly and effectively give the price of cloud resources,and on the other hand can provide a reference for domestic IT enterprises.更多还原