摘要云计算是一种将计算任务分布在大量计算机构成的资源池上的商业计算模型，用户 可以动态按需申请资源，云计算具有超大规模，虚拟化，高可靠性，通用性，按需服务 和极其廉价等优点。Hadoop 平台是云计算下最流行的分布式系统框架，可靠性好，同时 也易于扩展。Hadoop 的任务调度技术是此平台的关键性技术，任务调度策略对于整体作 业的执行效率至关重要，而遗传算法作为一种常用的启发式优化算法，能较好的计算出 最优的任务分配序列，适用于 Hadoop 的任务调度。本文设计实现了基于遗传算法的 Hadoop 平台任务调度算法，主要工作如下：

1）在介绍 Hadoop 的存储管理，MapReduce 计算框架，作业调度系统和遗传算法等概念的 基础上，给出了基于遗传算法的 Hadoop 任务分配模型描述，重点阐述了算法中若干关键 问题，包括：编码和解码、选择算子、交叉算子、变异算子、适应度函数、初始种群的 产生、收敛条件和其它约束条件等，并给出了算法的具体流程步骤。 2）搭建了实验平台，给出了算法中相关类的实现，并介绍了算法的加载，最后给出了实 验结果，并对其进行了分析。77728

毕业论文关键词 Hadoop 平台 任务调度 遗传算法

毕 业 设 计 说 明 书 外 文 摘 要

Title The task scheduling algorithm based on genetic algorithm

Abstract Cloud computing is a business computing model which put computing tasks in resource pool composed by a lot of computers,the user can dynamically apply resources as needed。Cloud computing has many advantages:large scale,virtualization,high reliability,versatility,on-demand service and very cheap。Hadoop platform is the most popular cloud computing distributed system framework,good reliability,but also easy to expand。Hadoop task scheduling technology is the key technology in the platform,task scheduling strategy is a key element to the efficiency of the overall job execution,and genetic algorithm as a commonly used heuristic optimization algorithm,used to calculate the optimal task assignment sequence,applicable to the Hadoop task scheduling。The task scheduling algorithm based on genetic algorithm is designed in this paper。The main work of this paper is as follows:

1)On the basis of the principle of Hadoop storage management,MapReduce computational framework,job scheduling systems and genetic algorithms,it gives the description of Hadoop task assignment model based on genetic algorithm,process steps of the algorithm and explains several key issues,including: encoding and decoding,selection operator,crossover operator,mutation operator,fitness function,generation of initial population and other convergence conditions。

2)set up the experimental platform, give the realization of related classes, and introduce the algorithm of loading, and finally get the results of the experiment and its analysis。

Key words  Hadoop platform  task scheduling  genetic algorithm

本科毕业设计说明书 I

目 次

1 绪论 1

1。1 研究背景 1

1。2 国内外研究现状 1

1。3 论文研究内容 2

1。4 论文结构 2

2Hadoop 平台介绍