基于多维度资源需求的改进遗传调度算法  

Improved Genetic Scheduling Algorithm Based on Multidimensional Resource Requirements

作　　者：高原 顾文杰[1,2,3] 彭晖 陈鹏[1,2,3] GAO Yuan;GU Wenjie;PENG Hui;CHEN Peng(NARI Group Corporation/State Grid Electric Power Research Institute,Nanjing 211106;NARI Technology Development Co.,Ltd.,Nanjing 211106;State Key Laboratory of Smart Grid Protection and Control,Nanjing 211106)

机构地区：[1]南瑞集团有限公司(国网电力科学研究院),南京211106 [2]国电南瑞科技股份有限公司,南京211106 [3]智能电网保护和运行控制国家重点实验室,南京211106

出　　处：《计算机与数字工程》2020年第4期753-758,785,共7页Computer & Digital Engineering

基　　金：国家重点研发计划项目(编号:2017YFB0902600);国家电网公司科技项目“大电网智能调度与安全预警关键技术研究及应用”(编号:SGJS0000DKJS1700840)资助。

摘　　要：针对传统的调度算法没有考虑磁盘IO,仅以执行时间为目标等不足,提出了一种基于多维度资源需求的改进遗传调度算法。算法引入IO资源作为输入参数之一,结合使用3种方法改进初始种群。算法对变异方法进行了优化,算法的适应度函数也采用多目标进行了设计。最后在模拟场景中进行了多次实验,验证了算法在任务负载均衡调度方面的性能较传统遗传算法有了显著提高。In view of the traditional scheduling algorithm does not take into account the disk IO,only aiming at execution time,an improved genetic scheduling algorithm based on multidimensional resource requirements is proposed.The algorithm introduces the IO resources as one of the input parameters,and improves the initial population combined with three methods.The algorithm also improves crossover and mutation in the process of task scheduling and fitness function is also optimized by using multiple objectives.Finally,a number of experiments are carried out in simulation scenarios to verify the effectiveness of the algorithm in the task load balance scheduling performance than traditional genetic algorithm has improved significantly.

关 键 词：资源管理 分布式 任务调度 初始种群优化 遗传算法 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]