云服务资源调度与市场交易模型研究

发布时间：2018-03-15 11:34

本文选题：云服务市场　切入点：优化调度　出处：《武汉理工大学》2015年博士论文　论文类型：学位论文

【摘要】：随着信息服务资源市场的不断扩大和需求信息服务用户的的逐日增多,作为信息服务商已经无法承担高额的服务成本,以往传统的服务资源市场组织管理方式已经无法满足云服务市场中用户多样化的需求服务要求,在这样的背景下促使了“云计算”服务的出现和大力发展。“云计算”作为一种新兴的信息资源服务形式,以基础设施、平台开发和软件应用的方式提供市场服务,把服务市场的信息资源从本地服务扩展到云服务市场环境中,并以快捷的方式提供给用户使用。云用户按照自身需求订购支付适合自己的个性化云服务,通过这种方式可以有效减少管理成本和投资成本。但是要获取这些服务的途径方式对于一些计算机专业人员来说,都是件费时费力的事情,那么对于大多数普通用户来说更是件不容易的事情,更不用说服务质量的有效保证。因此,高效优化的云服务市场机制使得云服务市场的每一个环节都井然有序是一个亟待解决的问题,具有十分重要的现实意义。针对上述问题利用市场经济优化理论、博弈理论、多目标优化等理论,结合云服务市场具体服务情况等相关知识和方法来寻求解决途径。首先从云服务市场的市场机制入手,分析了市场竞争机制、资源调度机制、交易机制、定价机制。随后依次研究构建了云服务资源优化调度模型、云服务资源定价模型、云服务资源交易模型。主要研究内容如下:(1)针对云服务市场中的资源调度优化问题,通过资源优化调度率选取与其高度相关的服务成本、服务时间、服务QoS属性值、用户的最优满意度,作为调度优化指标,构建了云服务资源优化调度模型,利用主客观赋权法根据用户的不同偏好属性对服务资源的属性进行赋权,更加科学贴切的反映用户意愿和资源调度的合理性。利用有序效用函数选取最优服务资源候选集,采用用户需求任务—服务资源—服务商的编码方式将任务、资源和服务商合理联系起来为更好的完成用户需求任务提供方法依据。最后通过模拟实验分析该模型的有效性。(2)制定云服务市场的定价模型。分别从以下三个方面着手进行:其一,根据用户效用最大化条件下,确定服务市场的合同和按需定价策略,并研究市场完全被垄断和引入竞争机制这两种市场机制下,此种定价方法的用户效用变化情况。其二,根据市场收益最大化,在合同和按需定价的基础之上确定市场最优定价策略。其三,根据市场的实际使用情况,在用户使用的某些时段会有大量的闲散资源存在,此时,需要研究制定现货实例定价策略来对这些闲散资源进行合理的利用。此时段的定价也被看做是一种调整用户市场服务使用行为的激励机制,从而实现市场均衡和整体市场的高效使用。(3)根据服务市场在高峰时段出现的拥挤现象,建立拥挤时段费用定价模型,首先根据优先权和限流费用来设定拥挤费用,然后从静态和动态两个角度去分析此拥挤收费在缓解市场压力方面的合理性,最后根据信息经济学中的博弈论,用户与用户之间围绕利益展开的非合作博弈,以及用户与云经纪人之间的非合作博弈,并结合上述优先权和限流收费政策,来解决云服务市场的拥挤现象。(4)针对服务资源市场的交易问题,本文构建了云服务市场交易体系该体系由云经纪人协调整个服务交易过程,进行用户需求任务的合理分配和市场资源的供需平衡控制。并提出了面向用户预算约束、时间约束以及预算和时间双重约束条件下的多实例组合交易决策模型。
[Abstract]:With the development of information service resource market continues to expand and the demand of information service user's daily increase as information service providers have been unable to bear the high cost of service, traditional service resources market organization management has been unable to meet the users of cloud services in the market demand of various service requirements, in this context the emergence of cloud "computing" services and develop. "Cloud computing" as a new form of information service, the infrastructure platform, software development and application way of providing marketing services, expand service market information resources from the local services to the cloud services market environment, and the quickest way available to users. Cloud users according to their own needs for personalized subscription payments cloud services of their own, by this way can effectively reduce management costs and investment costs. But To get the service mode for some computer professionals, is a time-consuming thing, so for most ordinary users, it is not an easy thing, not to guarantee service quality. Therefore, the cloud service efficient market mechanism optimization makes every link of the cloud services market in order is an urgent problem, has very important practical significance. In order to solve the above problems by using market economy optimization theory, game theory, optimization theory, combined with the knowledge and methods of cloud services and other related market specific services to seek solutions. Starting from the market mechanism of cloud services market, market analysis the mechanism of competition, resource scheduling mechanism, trading mechanism, pricing mechanism. Then the research constructs the optimization model of cloud services resources, cloud service resource pricing Model of cloud service resource transaction model. The main contents are as follows: (1) to solve the optimization problem of resource scheduling of cloud services in the market, through the resource optimization scheduling rate selection and service cost, highly relevant Business Hours, QoS attribute values, optimal user satisfaction, as optimization index, construct the optimal scheduling model of cloud the service resources, using subjective and objective weighting method according to different attributes preference attributes of users of the service resources to weight more scientific and appropriate to reflect the rationality of user intention and resource scheduling. Using the ordered utility function to select the optimal resource service candidate set by user needs task service resources - Service providers encoding tasks and resources reasonable link and service providers can provide a theoretical basis for better user needs to complete tasks. Finally, through the simulation and analysis of the effectiveness of the model (2). The development of cloud services market pricing model. Respectively from the following three aspects: first, according to the user's utility maximization under the condition of market, determine the service contract and on-demand pricing strategy, and study the market is completely monopoly and introduce the competition mechanism of the two kinds of market mechanism, change the user utility pricing method secondly, according to the market to maximize profits in the contract, and to determine the optimal pricing strategy according to market demand pricing basis. Thirdly, according to the actual situation of the market, in some time users will have a large number of idle resources, at the same time, the need to develop research spot pricing strategy to these examples of idle resources reasonably the use of incentive mechanism. At this time the price is also seen as a service market to adjust the user behavior, so as to realize the efficient use of market equilibrium and the overall market (3) root. According to the congestion service market appeared in the peak period, a congestion cost pricing model according to the first time, priority and limiting costs to set the congestion cost, and then from two angles of the static and dynamic analysis of the congestion pricing rationality in relieving the pressure of the market, according to the game theory in the information economics theory, all around the interests the non cooperative game between users and between users and cloud broker non cooperative game, and combined with the priority and limiting the charging policy, to solve the congestion phenomenon of cloud services market. (4) aiming at the problem of transaction service resource in the market, this paper constructs the cloud services market transaction system the system consists of cloud broker coordinate the whole process of service transactions, the supply and demand balance control reasonable allocation of user demand tasks and market resources. And put forward the user oriented budget constraint, time A multi instance portfolio decision model under the dual constraints of budgetary and time constraints.

【学位授予单位】：武汉理工大学
【学位级别】：博士
【学位授予年份】：2015
【分类号】：F49

【参考文献】