| Summary: | 碩士 === 國立中央大學 === 資訊工程研究所 === 99 === Thanks to advances in wide-area network technologies and the low cost of computing resources, grid computing came into being an active research area. One motivation of grid computing is to aggregate the power of widely distributed resources, and to provide non-trivial services to the users. To minimize the total completion time (makespan), an efficient grid scheduling mechanism must be used in a grid system to dispatch computing tasks to computing resources effectively. However, it has been proved that the optimal scheduling algorithm is NP-hard. Therefore, many people turn to use heuristic approaches for grid scheduling. In this thesis, we introduce eleven common scheduling heuristics to schedule a combination of linear dependent jobs and independent jobs. Then, we use a grid simulator, namely GridSim[1], to evaluate the performance of these heuristic approaches. According to the simulation results, we propose a novel hybrid heuristic approach that can avoid the drawbacks of the eleven heuristic approaches under different situations. Further simulation results confirm that the proposed hybrid approach is among the best heuristic approaches under most circumstances.
|
|---|
