Abstract: For the distributed resources sharing of grid, an improved ant colony algorithm is proposed to apply to grid resource scheduling. First, trust value is applied to the algorithm as the important pheromone of the ant remnants. Through the evaluation of the trust value, the algorithm can effectively improve the search ability and the reliability of grid resource. Secondly, a local and global pheromone updating mechanism is designed to realize load balance at each resource node. Finally, GridSim is used to simulate the strategy of campus grid resource scheduling based on the improved ant colony algorithm by choosing some appropriate parameters. Experiment shows the that, with the aid of the algorithm, the utilization rate of grid resources and success rate of task committing is increased, and the execution time of grid resource scheduling is reduced , the performance of the grid system is thereafter improved.

Key words: grid resource scheduling, improved ant colony algorithm, trust value, pheromone, gridsim