基于蚁群优化算法的网格任务映射策略

Tasks mapping in grid computing environment based on ACO algorithm

针对网格环境下实现任务最优映射的问题,提出一种基于蚁群优化算法的网格任务映射策略(ACO-GTM)。该算法通过人工蚂蚁在构建图上行走构建初始解,利用最优改进2-选择局部搜索方法对初始解进行局部优化,并采用全局信息素更新与局部信息素更新相结合的信息素更新策略。最后通过实验与其他算法进行比较,表明所提出的映射算法在最优跨度和负载平衡方面具有明显的优越性。

In order to optimize the tasks mapping in grid, a grid tasks mapping algorithm based on Ant Colony Optimization （named ACO-GTM） was proposed. The algorithm generated initial solutions through these artificial ants traversed on the construction graph and optimized these initial solutions by using the Best-improvement 2-opt local search algorithm. It combined the global and local pheromone updates. The experiments show that the proposed algorithm for the mapping problem has better performance than other algorithms on optimum makespan and load-balancing.