对等网络环境下多目标任务容错调度方法研究  

Research on Fault Tolerant Scheduling Method of Multi-Objective Tasks in Peer-to-Peer Network Environment

作　　者：秦轶翚 马涛 QIN Yi-hui;MA Tao(Beijing Union University,Beijing 100011,China)

机构地区：[1]北京联合大学,北京100011

出　　处：《计算机仿真》2021年第8期352-355,共4页Computer Simulation

摘　　要：为了加强网络环境的可靠性、强化系统运算性能,提出对等网络环境下多目标任务容错调度方法。使用PB算法对处理器进行容错处理,划分任务主、副版本,通过分析任务相应版本与所在处理器的运行时间,构建任务模型与故障模型。拟定多目标任务不同开始时间与时限,通过自适应策略拟定启发式多目标任务容错分配策略,将多目标任务转换为回路任务。设定回路任务的数量与主版、副版本在普通处理器内的运行时间,判断调度序列内任务,计算其在处理器内需要消耗的时间,将结果对比之前拟定的时间判断是否调度成功,成功则启动副版本迭代计算。实验结果表明,所提方法不会因为网络节点数量不同,出现大量任务调度失败的问题,提高了系统性能和网络环境安全。This paper proposes a fault-tolerant scheduling method for multi-objective tasks in a peer-to-peer network environment for improving the reliability of the network environment and enhancing the computing ability of the system.PB algorithm was applied to a fault-tolerant processor.The main and sub versions of the task were divided,and the corresponding version of the task and the running time of the processor were analyzed to build the task model and fault model.Different start times and time limits of multi-objective tasks were formulated.According to the adaptive strategy,a heuristic fault-tolerant assignment strategy for multi-objective tasks was proposed.The multi-objective task was transformed into a loop task.The number of loop tasks was set,and the running time of the main and sub versions in the common processor was also set.Tasks in the sequence were judged and scheduled,and the time consumed in the processor was also calculated.The calculation results were compared with the planned time.If the scheduling is successful,the sub version will be started to iterate,and vice versa.The experimental results show that the method has excellent system performance and high network security,and is not easily affected by the number of different network nodes.

关 键 词：对等网络 多目标任务 容错调度 故障模型 任务模型 

分 类 号：TP316[自动化与计算机技术—计算机软件与理论]