网格任务调度优化模型和算法分析-optimization model and algorithm analysis of grid task scheduling.docxVIP

II II ABSTRACT With the rapid development in scientific technology，the calculation quantity of a large amount of scientific computing and complicated applications has been increasing dramatically. As one kind of the significantly developing infrastructure, the Grid Computing can share a large-scale of calculation resource, storage resource, data resource, software resource, equipment resource, HR and so on. The Grid Computing can overcome the existing calculating limitation and realize the large-scale cooperative scientific computing and cooperatively solve problems, which provides a novel computing mode to solve some very large-scale, super-complex computing-intensive or data-intensive problems. The grid research is a great challenge because its resources are characterized by the large-scale distribution, heterogeneous types of diversity and dynamic variability. Task scheduling is one of hotspots in the current Grid Computing research. The thesis studies the task scheduling and load proportion, and we have done the following works: Make a comprehensive summary about the research background and status of grid computing in the United States, Europe, Japan and China. Investigate the definition of grid computing and its characteristics，applications and research classification. Analyze three system structure of grid, such as the five-layer sandglass architecture, the open grid service architecture and WEB service resource architecture. Analyze the basic characteristic and target of the task scheduling which were used in grid computing and make the formalized definition of the task scheduling. Study three situations of the load balancing strategies which were used for achieving the balanced distribution of tasks load among resource nodes in grid computing. Study three intelligent algorithms including the genetic algorithm, the immune clone algorithm and the simulated annealing algorithm which could realize the task scheduling and load balancing effectively. The