• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

J4 ›› 2007, Vol. 29 ›› Issue (3): 63-65.

• 论文 • 上一篇    下一篇

基于蚁群算法的分布多处理机系统任务调度研究

孔晓红[1,2] 须文波[1] 孙俊[1]   

  • 出版日期:2007-03-01 发布日期:2010-05-30

  • Online:2007-03-01 Published:2010-05-30

摘要:

本文主要基于现代蚁群算法讨论分布式系统调度。蚁群算法是一种构造型启发算法,在离散优化问题中得到广泛应用。分布式系统调度属于NP-hard,为了提高算法性能,把问题任务图的优先级作为启发信息。最后,采用随机产生的任务图将调度结果和模拟退火算法、遗传算法等进行了比较。

关键词: 蚁群算法 模拟退火 分布式系统调度 优先级

Abstract:

This paper introduces the ant colony algorithm to solve the distributed system scheduling problem with arbitrary task graphs.The ant colony algorithm   is a novel constructive heuristic approach to optimizing discrete space functions.Distributed system scheduling is a NP-hard problem.In this paper we pr esent a problem priority-based heuristic method,which can enhance algorithm performance.The results for arbitrary task graphs are compared to simulate t he annealing algorithm and other algorithms.

Key words: ant colony algorithm;simulated annealing;multiprocessor scheduling;genetic algorithm