الفهرس | Only 14 pages are availabe for public view |
Abstract Grid computing is a high performance computing environment to solve large-scale computational demands. Computational grids are emerged as a next generation computing platform which is a collection of heterogeneous computing resources connected by a network across dynamic and geographically dispersed organizations. This is carried out to form a distributed high performance computing infrastructure. In computational grid, the main emphasis is given on resource management and job scheduling. The main goal of scheduling is to minimize the processing time of jobs. Various research efforts have been performed on job scheduling problem in grid, but still further analysis and research needed to be done to improve the performance of scheduling algorithm in computational grid. In this thesis, Fast Artificial Fish Swarm Algorithm has been proposed for job scheduling in computational grids. The basic idea of FAFSA is to merge the Levy flight algorithm for randomization in the original FSA. FSA is algorithm imitate the fish behaviors such as preying, swarming, and following with local search of an individual fish for reaching the global optimum. We used levy random algorithm to replace normal fish random behavior, the levy random algorithm adds enhancement in FSA algorithm and enhances fish random behavior that finally effect in FSA performance. Beginning in this thesis, there are many of measuring performance in job scheduling like makespan, flow time, Job lateness and Job tardiness. We have specifically focus on improving computational grid performance in terms of makespan to minimize the average completion time of jobs through optimal job allocation on each grid node in application-level scheduling. A simulation of proposed approach using GridSim toolkit is conducted. A comparison of our proposed approach with FSA job scheduling strategies is provided. Experimental results show that our proposed algorithm performs efficiently in computational grid environment and we conclude that the proposed modification in FSA provides better performance than the original FSA, the levy random algorithm adds enhancement in FSA algorithm and enhances fish random behavior that finally effect in FSA performance improvement in all algorithms is between 4-5%. |