| Summary: | 碩士 === 中華大學 === 資訊工程學系(所) === 96 === Grid computing can integrate computational resources from different networks or regional areas into a high performance computational platform. With the use of this high performance platform, complex computing-intensive problems can be solved efficiently. Scheduling problem is an important issue in a grid computing environment. Because of the differences in computational capabilities and network status of computational resources, an efficient scheduling algorithm is necessary to assign jobs to the appropriate computing nodes. In this thesis, we propose two dynamic scheduling algorithms GDSA and EDSA for scheduling tasks in grid computing environment. The proposed algorithms use the optimal-searching technique of genetic algorithm (GA) to get an efficient scheduling solution in grid computing environment and adapt to different number of computing nodes which have different computational capabilities. And, two types of chromosomes were used to discuss the effect on performance. Furthermore, the hybrid crossover and incremental mutation operations within the EDSA algorithm can move the solution away from the local-optimal solution towards a near-optimal solution. In order to verify the performance of the algorithms, a simulation with randomly generated task sets was performed, and they were then compared with five other scheduling algorithms. The simulation results show that the use of GA can effectively evolve a better schedule than other conventional scheduling algorithms. Especially, the proposed EDSA outperformed among all other scheduling algorithms across a range of scenarios.
|
|---|
