基于多目标优化算法的云计算调度任务分配及优化  

Research on task allocation and optimization of cloud computing scheduling based on multi objective optimization algorithm

作　　者：沈建国[1] SHEN Jianguo(Wuxi Institute of Business and Technology,School of Internet of Things and Artificial Intelligence,Wuxi214000,China)

机构地区：[1]无锡商业职业技术学院物联网与人工智能学院,江苏无锡214000

出　　处：《延边大学学报(自然科学版)》2025年第1期19-24,共6页Journal of Yanbian University(Natural Science Edition)

基　　金：中央高校基本科研业务费专项资金资助项目(020414380195)。

摘　　要：为了优化计算机对队列任务的调度效率,将双深度Q网络和改进鲸鱼优化算法相结合提出了一种基于多目标决策优化的任务调度模型.在该模型中,双深度Q网络主要用于拟合计算机的任务调度过程,以实现更加高效的任务分配与优化.改进鲸鱼优化算法用于协同分配生成的任务队列,以提升任务分配的整体效率,研究结果显示,与蚁群算法、粒子群优化算法、灰狼优化算法相比,该调度模型在迭代过程中的收敛速率最高,最终任务执行时间为182ms,且任务执行效率比其他算法提升了15.71%~31.34%.另外,在低任务量状态下,该模型的调度效果也明显优于其他对比算法,且任务逾期时间范围不超过10ms.上述结果表明,该模型经过优化后能够保持较好的任务调度效果,且可有效提升计算机对服务器资源的利用率,因此该算法可为计算机的任务调度提供良好参考。In order to optimize the scheduling efficiency of queue tasks by computers,a task scheduling model based on multi-objective decision optimization is proposed by combining dual deep Q-networks and improved whale optimization algorithm.In this model,the dual deep Q-network is mainly used to fit the task scheduling process of computers,in order to achieve more efficient task allocation and optimization.An improved the whale optimization algorithm is used for collaborative allocation of generated task queues to enhance the overall efficiency of task allocation.The research results show that compared with ant colony algorithm,particle swarm optimization algorithm,and grey wolf optimization algorithm,this scheduling model has the highest convergence rate during the iteration process,with a final task execution time of 182 ms,and a task execution efficiency improvement of 15.71%~31.34%compared to other algorithms.In addition,under low workload conditions,the scheduling performance of this model is significantly better than other comparative algorithms,and the task overdue time range does not exceed 10 ms.The above results indicate that the optimized model can maintain good task scheduling performance and effectively,improve the utilization of server resources by computers.Therefore,this algorithm can provide a good reference for computer task scheduling.

关 键 词：多目标决策优化 双深度Q网络 改进鲸鱼优化算法 高斯收敛 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]