Research On The Improved Grid Schedule Algorithm Based On Min-min And Max-min

Task scheduling is to achieve high-performance grid computing a basic problem, scheduling strategies directly affect the performance of the grid, the grid's own distribution, heterogeneous, dynamic, and autonomous and other characteristics of the traditional scheduling algorithm proposed new challenge, grid task scheduling algorithm to become a hot topic in grid research.In this paper, the concept of grid, grid separation, grid architecture, and other related issues were introduced, and then task scheduling in grid computing objectives, principles and basic strategies are discussed. For the heterogeneity of the grid in order to balance the goal of task scheduling, presents a heterogeneous systems to quickly balance the grid task scheduling algorithm, the algorithm uses the greedy thought, first of all tasks to be assigned to a unit on average each machine, and then find out with the maximum load and minimum load of the machine, transfer or exchange them to the above tasks, so that the two machines on the task of rapid equilibrium, and then through repeated iterations, so that the grid system, each machine the task of quickly reach equilibrium. Theory approach used to prove the correctness of the algorithm, and then Min-Min, Max-Min algorithm evaluation benchmark, the use of Grid Scheduling Simulation Toolkit GridSimToolkit carried out extensive simulation results show that: FILT algorithm has good performance , can be compared with Min-Min, Max-Min algorithm is better scheduling results.In addition to the final paper summarizing the research work, but also for future research directions are also prospected.