论文部分内容阅读
针对异构网格资源下任务的调度最小化执行时间问题(NP难题),提出了一种基于云模型的自适应蚁群调度算法.该算法在定性知识的指导下,权衡提高收敛速度和保持解的多样性之间的矛盾,能够自适应控制搜索范围,较好地避免了传统蚁群算法易陷入局部最优解和选择压力过大造成的早熟收敛等问题,提高其快速寻优能力.实验结果表明该算法在保证有效的加速比的同时具有精度高、收敛速度快等优点,极大地提高了网格任务调度的规模和效率.
Aiming at the problem of minimizing the execution time of tasks in heterogeneous grid resources (NPs), an adaptive ant colony scheduling algorithm based on cloud model is proposed. This algorithm, under the guidance of qualitative knowledge, improves the convergence speed and keeps Which can adaptively control the search range and avoid the problems that the traditional ant colony algorithm is apt to fall into the local optimal solution and the precocious convergence caused by the over-selective pressure so as to improve its rapid optimization ability. The experimental results show that the proposed algorithm has the advantages of high precision and fast convergence while guaranteeing effective speedup, which greatly improves the scale and efficiency of grid task scheduling.