异构计算系统中基于遗传算法的任务分配与调度  

Task Matching and Scheduling by Using Genetic Algorithms in Heterogeneous Computing System

作　　者：张聪[1] 马义忠[1] 

机构地区：[1]兰州大学信息科学与工程学院,兰州730000

出　　处：《微电子学与计算机》2004年第6期74-78,189,共6页Microelectronics & Computer

摘　　要：任务分配与调度是异构计算系统(HCS)的核心问题之一。本文提出了一种对异构计算系统进行任务分配与调度的遗传算法。首先对HCS任务分配调度问题做出定义,然后分别讨论了问题的编码表示,种群初始化,染色体适应度函数设计,杂交、内部杂交与迁移遗传算子设计,控制参数自适应,算法的总体框架等。最后给出算法的仿真实验情况及实验结果分析与结论等。Task matching and scheduling is one of the important problems in HCS (Heterogeneous Computing System). In this paper,we propose a genetic algorithm for task matching and scheduling in HCS.First the task matching and scheduling problem for HCS is described. Then the permutation representation, the initialization of population,the design of fitness function,the crossover, internal crossover and migration operator design, the adaptive control parameter, the structure of the algorithm, are discussed respectively. In the end, the algorithm simulation, experiment result analysis and the conclusions are presented.

关 键 词：异构计算 任务分配与调度 遗传算法 

分 类 号：TP314[自动化与计算机技术—计算机软件与理论]