面向多等级应用的气象云资源调度方法研究  

Research on Resource Scheduling in Meteorological Cloud Environment for Multi-class Application

作　　者：梁中军 孙志于[1] 韩同欣[2] 宋雅婷[1] 张正阳[1] LIANG Zhong-jun;SUN Zhi-yu;HAN Tong-xin;SONG Ya-ting;ZHANG Zheng-yang(Xinjiang Meteorological Information Center,Urumqi 830002,China;National Meteorological Information Center,Beijing 100081,China)

机构地区：[1]新疆气象信息中心,新疆乌鲁木齐830002 [2]国家气象信息中心,北京100081

出　　处：《计算机技术与发展》2022年第8期203-209,共7页Computer Technology and Development

基　　金：国家重点研发计划课题(2018YFC1507102,2019YFA0606904)。

摘　　要：气象云平台是集约支撑气象业务发展基础资源平台,如何合理调度气象云平台的基础资源已成为气象业务稳定运行的关键。由于不同等级气象应用在汛期和日常具有特殊支撑需求,以及部分功能组件对气象云平台的物理服务器有约束要求,现有调度算法难以被直接应用。为此,该文对问题建模,设计目标函数分别评价气象云平台的资源利用水平和对重要核心业务的支撑情况,基于物理服务器与功能组件之间的约束关系,将问题建模成一个多约束多目标优化问题。然后,设计一种改进蛙跳算法的气象云资源调度方法求解模型,该方法利用适应度函数来综合评价模型目标,并重新定义蛙跳算子,利用局部最优交叉操作和最优青蛙变异策略来迭代搜索最优资源调度方案。最后,通过实验验证了该方法的效果。Meteorological cloud platform is a basic resource platform that supports the development of meteorological services intensively.How to reasonably schedule the basic resources of meteorological cloud platform has become the key to the stable operation of meteorological services.Because different levels of meteorological applications have special support requirements in flood season and daily life,and some functional components have constraints on physical servers of meteorological cloud platform,existing scheduling algorithms are difficult to be directly applied.Therefore,we model the problem and design objective functions to evaluate respectively the resource utilization level of meteorological cloud platform and its support for important core business.Based on the constraint relationship between physical server and functional components,the problem is modeled as a multi-constraint multi-objective optimization problem.Then,a weather cloud resource scheduling model with improved leaping frog algorithm is designed.The fitness function is used to comprehensively evaluate the model objective,and the leaping frog operator is redefined,and the local optimal crossover operation and the optimal frog mutation strategy are used to iteratively search for the optimal resource scheduling scheme.Finally,the effectiveness of the proposed method is verified by experiments.

关 键 词：气象云 资源调度 多等级应用 蛙跳算法 多目标优化 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]