Research On Task Scheduling In Grid Environment For Power Marketing System

Grid technology is a new IT technology following the Internet and the Web, also a development direction of parallel and distributed processing technology. Grid task scheduling technique is an important aspect of grid study, which according to the characteristic of tasks with appropriate scheduling strategy distributes the tasks to different resources nodes to run. Because of the distribution and heterogeneous of the grid resources, as well as a variety of different needs of grid resources, the task scheduling becomes extremely complicated, and the task scheduling algorithm directly affects the efficiency of the execution of task in grid environment. At present, most task scheduling strategies are proposed for independent tasks, but in practice, a large application task often require to multiple into with certain sub-constraints tasks, as related tasks. Then follow a certain precedence relations and use the associated task scheduling strategies to do scheduling, which is the contents of this paper.This paper studies the task scheduling strategy for the power marketing system in grid environment. First using task decomposition technique to decompose the application of the loosely coupled type, divide the large application task into sub- tasks with priority topology. Then according to the task features of power marketing system, we propose the corresponding task scheduling strategy. The main studies on the ant colony algorithm and improvement in the grid task scheduling strategy, and applied to the associated task scheduling. Finally, do experiment in the GridSim of grid simulation tool for the ant colony algorithm. The results show that the improved algorithm compared with basic ant colony algorithm, significantly reducing the overall task execution time, and improving the performance of grid task scheduling, is an efficient and feasible scheduling algorithm.