摘要
随着云计算技术的快速发展以及用户需求的增长,资源分配效率不能满足时间需求,服务质量同步下降。因此,研究提出了基于遗传算法改进量子粒子群优化算法的云计算资源调度算法,融合算法前半段采用遗传算法进行迭代,扩大搜索范围,后半段使用量子粒子群算法,改进收缩扩张系数,利用OpenStack构建资源调度平台。实验表明,融合算法花费的时间分别比其他算法低了2.6 s以及4.1 s,能耗分别比其他两种算法低0.11和0.14,负载率熵最低,分别比其他3种算法低0.09、0.03以及0.10。融合算法的CPU使用量最大预测误差分别比其他算法低了12.1和1.3,平均内存预测误差分别比其他算法低了1.7和1.4。由此可得,融合算法能够有效加快资源调度速度,提升资源利用率和节点负载均衡性。
With the rapid development of cloud computing technology and the growth of user demand,the efficiency of resource allocation cannot meet the time demand,and the quality of service declines simultaneously.Therefore,this paper proposes a cloud computing resource scheduling algorithm based on genetic algorithm to improve quantum particle swarm optimization algorithm.In the first half of the fusion algorithm,genetic algorithm is used to iterate and expand the search scope;in the second half,quantum particle swarm algorithm is used to improve the contraction and expansion coefficient,and OpenStack is used to build a resource scheduling platform.Experiments show that the fusion algorithm takes 2.6 s and 4.1 s less time than the other algorithms,the energy consumption is 0.11 and 0.14 less than the other two algorithms,and the load rate entropy is the lowest,which is 0.09,0.03 and 0.10 less than the other three algorithms,respectively.The maximum prediction error of CPU usage of fusion algorithm is 12.1 and 1.3 lower than that of other algorithms,and the average memory prediction error is 1.7 and 1.4 lower than that of other algorithms.It can be seen that the fusion algorithm can effectively speed up resource scheduling,improve resource utilization and node load balancing.
作者
宋志伟
SONG Zhiwei(Guangzhou Vocational and Technical University of Science and Technology,Guangzhou 510555,China)
出处
《自动化与仪器仪表》
2025年第7期84-87,92,共5页
Automation & Instrumentation
基金
2021年度广东省普通高校青年创新人才类项目《基于云计算的智能工程系统设计与开发》(2021KQNCX146)。
关键词
遗传算法
粒子群算法
云计算
资源调度
OPENSTACK
genetic algorithm
particle swarm optimization
cloud computing
resource scheduling
OpenStack
