Research On Grid Task Scheduling Based On Hybrid Ant Colony Algorithm

Grid task scheduling is an important part of the grid, which directly affects the performance of grid computing system. But the gird task scheduling has the problems of needing a long time for task scheduling or load imbalance of the resources. Ant colony algorithm is incorporated into grid task scheduling to overcome the shortcoming mentioned above, and the performance of grid will be improved.Based on the overall study of the theory and performance of the grid task scheduling, ant colony algorithm and tabu search, in this paper, an algorithm for grid task scheduling based on hybrid ant colony algorithm and Mathematical model are proposed. The algorithm uses tabu search as the local search of ACO, those optimization solutions is generated by ACO, then tabu search expands the search space of the solutions based on the optimization solutions. The algorithm overcomes Tabu search's dependence on the initial solution and makes ACO avoid falling into local search by using the characteristics of tabu search; furthermore, in order to solve the problems of grid fault—tolerant, network delay and load balance, the paper improves the composition and the updates rules of the pheromone to shorten the time of grid task scheduling and the load balance of the resources.To evaluate the performance of the the gird task scheduling algorithm, this papers simulate the algorithm with the GridSim Toolkit. The experiment results show that the algorithm can effectively shorten the time of the task scheduling and improves the load balancing of the resources.