Research On Grid Resource Scheduling Model Based On Genetic Algorithm

Grid Computing, which has been called the next generation of the Internet, has been focused from the day of its born. It's just perfect as a super computer. According to the Grid, we could incorporate not only the computer and website, but also other information, for example: the database, the software and other various equipment of information obtained, even, we can use the electric equipment that connects to the Grid.When we study and research the Grid Computing, no doubt, the work of resource scheduling is the most important. Only when we do with the scheduling problem better, design the more efficient and reasonable scheduling model, we can show the advantage of Grid Computing. In the system of resource scheduling, there are some scheduling algorithms, for instance: the Min-min, the Max-min, Fast-Greedy and so on, but in fact, these algorithms sometimes occur some errors and problems, leading to the scheduling could not reach the best or could not be fastest to reach. For the sake of this status, the research of resource scheduling that introduce to a efficient algorithm emerges as the times required.In this article, we use the GA as the algorithm of the resource scheduling, the GA, as one cute, simple algorithm, has showed evident advantage to make as perfect as possible. According to uniting the resource scheduling and the GA, make the resource change to the chromosome of the GA, so that we can find the most effective result. Before running the GA, we recount the building of the Grid environment detailedly, this is a necessary step, then we refered the steps of Grid exploitation of this model, including service interface, some important files that associate the Grid, implementing the service and deploying the service and so on. These job is the key work that could fuse the GA. Finally, we use the emulator and experiments, we showed the result upon two groups of data, proved that the GA suit the task scheduling of big size, when the size of tasks become bigger and bigger, the convergent time of the GA declined distinct, and efficiency is improved. Besides, we showed a compared data result between the GA and the traditional algorithms of the Grid resource scheduling called Min-min, proved that the GA indeed better than the traditional algorithms of resource scheduling in the executing speed and the efficiency.