移动云计算环境下任务分配问题的研究

发布时间：2018-11-23 15:23

【摘要】：随着网格计算、云计算等新技术的不断涌现和发展，怎样在计算系统内各个节点合理分配任务以达到资源的合理利用和负载均衡的目的成为了研究的热点。移动云计算作为云计算的一种延伸和拓展，面临着设备差异化大、移动网络情况复杂且开销较大等问题，而移动云计算环境下的任务分配研究则相对简单滞后。云端增效模式是目前最常见的移动云计算模式。云端增效模式下的应用模型根据任务分配方式可以分为固定模型和弹性模型，其中，弹性应用模型可以根据移动设备的状态和带宽弹性地将任务在云端和移动端分配，相比于固定分配的应用模型，能有效避免移动设备资源的浪费、减少网络开销。论文借鉴了双处理器任务分配问题的研究方法，提出了基于最大流理论的弹性任务分配算法，将任务分配转化为网络流的求最小割问题，并根据弹性模型重新设计了网络流的构建方式、容量的计算公式等具体步骤，同时引入权限值将多目标优化问题转换为单目标优化问题。算例分析表明该算法能够减少应用执行时间、降低网络费用。微云模式是未来移动云计算发展的热点，与云端增效模式不同的是，其节点数目和节点计算能力各异，也没有中心节点专门负责任务分配，各个节点只能根据自身情况和系统的反馈进行任务分配，，启发式算法可以作为解决这类问题的有效算法。论文借鉴了网格计算中任务分配问题的研究方法，将蚁群算法应用于微云模式的任务分配中，并给出了方法的步骤。仿真实验表明，该方法相比随机分配执行时间更短、负载也更加均衡。不同模式下，采用合理的任务分配方法能够减少移动应用的运行时间和流量费用，为用户带来更好的使用体验。
[Abstract]:With the continuous emergence and development of new technologies such as grid computing cloud computing and so on how to allocate tasks reasonably in the computing system to achieve the purpose of rational utilization of resources and load balancing has become a hot research topic. As an extension and extension of cloud computing, mobile cloud computing is faced with such problems as wide differentiation of devices, complex and expensive mobile networks, while task allocation research in mobile cloud computing environment is relatively simple and lagging behind. Cloud efficiency model is the most common mobile cloud computing model. The application model in cloud efficiency mode can be divided into fixed model and elastic model according to the task allocation mode, in which the elastic application model can flexibly distribute tasks in the cloud and mobile according to the state and bandwidth of the mobile device. Compared with the fixed allocation model, it can effectively avoid the waste of mobile device resources and reduce network overhead. In this paper, a new elastic task allocation algorithm based on the maximum flow theory is proposed, which can transform the task assignment into the minimum cut problem of network flow. According to the elastic model, the construction method of network flow and the calculation formula of capacity are redesigned. At the same time, the multi-objective optimization problem is transformed into a single-objective optimization problem by introducing the permission value. The example shows that the algorithm can reduce the application execution time and the network cost. The micro-cloud model is the hot spot of mobile cloud computing in the future. Unlike the cloud synergy model, the number of nodes and the computing power of nodes are different, and no central node is specifically responsible for task allocation. Each node can only assign tasks according to its own situation and the feedback of the system. Heuristic algorithm can be used as an effective algorithm to solve this kind of problem. Based on the research method of task assignment in grid computing, the ant colony algorithm is applied to the task allocation of micro-cloud model, and the steps of the method are given. Simulation results show that the proposed method has shorter execution time and more load balance than random allocation. In different modes, reasonable task allocation method can reduce the running time and traffic cost of mobile application, and bring users a better experience.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.07

【相似文献】