Research Of Grid Task Scheduling Based On Genetic Algorithm

Task scheduling algorithm is one of the most important parts in grid research. How allocate tasks to different resources in order to make the grid system obtain the highest performance is the problem which the task scheduling algorithm needed to resolve. Because of the isomeric and the dynamic of the grid, and the different resource requirements of the application program, the grid task scheduling becomes more complex. A bad task schedule strategy may increase the runtime of the tasks, and descend the throughput of the grid system.First, this paper introduces the development of the grid. It analyzes the characters of the task scheduling, and concludes the influences of the task scheduling algorithm. One mathematical model is given out, based on the research. A scheduling algorithm based on genetic algorithm is addressed. In order to ensure the diversity of population and the global search capability of GA, a new primary population forming method is addressed base on the essential and sufficient conditions of the schedule list's legality. Adaptive genetic operators are introduced, and according to the value of individual fitness different crossover and mutation operations are performed, aims at the characters of chromosome's structure. It ensures globe convergence while keeping diversity of population. At last, we've made experiments by using the SimGrid, and the algorithm's efficiency has been proved.The grid task scheduling problem is a typical NP problem. Applies genetic algorithm to the grid task scheduling problem, can decrease the total runtime of the tasks and raise the usage of the resources.