摘要
分布式计算系统中的一个根本问题是任务模块在处理器上的合理分配，以使总费用最小。针对随机试探法对初始条件敏感的不足，本文利用改进的遗传算法，通过设计合理的遗传算子寻求该任务分配问题的最优解。实验结果表明，本文的方法对初始条件不敏感，对具有不同拓扑结构的一致性及非一致性任务分配问题，其平均总费用降低约２％ ，此外，在大多数情况下也能使完成费用降低。
It is a fundamental task to assign interacting task modules to heterogeneous processors in a distributed computing system. An improved genetic algorithm is proposed to solve the problem so that the total cost incurred is minimized. Expe rimental results show that, in addition to its insensitiveness to the initial states, the method proposed here can provide major improvement over the stochastic probe method for both the uniform and non uniform task assignment problems. Furthermore, in most cases, the average completion cost is less than that obtained by the stochastic probe method.
出处
《数据采集与处理》
CSCD
1999年第3期293-297,共5页
Journal of Data Acquisition and Processing
关键词
任务分配
遗传算法
随机试探法
分布式计算机
distributed systems
task assignment
genetic algorithms
stochastic probe
