Research On Grid Task Scheduling Algorithm Based On Heuristic Method

Grid can integrate various resources which are broadly distributed everywhere, and then it can afford the computing ability to solve the problems which are very complex and in need of a great deal of recourses. Real-time application under grid environment is the extension of grid technology in real-time application field. It is a new research field of grid technology. The complexity of grid and the variety of the recourses on grid make the real-time task scheduling much harder. The key point is how to take all characters of grid into consideration when scheduling. The traditional task scheduling method is not suitable for grid.Heuristic method is a good way to solve these problems. Analyze the design principle heuristic method, the characteristics of grid and the real-time tasks, and then illustrate usability of heuristic method in soft real-time task scheduling under grid. After that, improve traditional heuristic real-time task scheduling in multiprocessors systems, bring it into grid real-time task scheduling, and then present a low time complexity e grid heuristic scheduling algorithm. This algorithm is a search process of the complete scheduling of the tasks. It ensures the task deadline by obligating recourses, reduces the time complexity by cutting down the search area and acquires the correct solution by taking many factors into consideration when setting up the heuristic function. Setting up the heuristic function also takes the communication spending between the processors and the recourses into consideration when scheduling. So the algorithm is very suitable for grid environment.The algorithm has piratical value in both theoretic and practice field. It reduces the time complexity and enhances the scheduling success ratio.