摘要
针对智能建筑室内环境下并行计算的动态任务调度问题,构建了基于分布式CPS思想的无线传感器网络(WSN)模型,并分别设计了基于可计算复杂性的任务分配策略和基于动态调度算法的任务调度策略。通过先将任务分配成若干个子任务,采用多带图灵机输入任务,由合适的计算节点进行计算,形成有向无环图,再按调度优先级排列任务,形成任务调度序列表,依序处理任务,从而达到了将任务分配、调度和执行相结合的目的。实验结果表明该策略可有效减少智能建筑室内环境分布式可计算WSN分布运行时任务之间的通讯时间和等待时间,同时提高了任务调度的成功率,最终优化系统的运行效率。
To solve the dynamic task scheduling problems of distribution parallel computing in intelligent building,a structure model of WSN based on distributed CPS conception is proposed. The task allocation strategy based on the com-putability complexity and dynamic scheduling algorithm based on the task scheduling strategy are designed. Firstly,the task is decomposed to a number of sub-tasks,Multi-band Turing machine is applied to the input of the task. The directed acyclic graph is formed though the calculation of the appropriate selected nodes. Second,task scheduling sequence tables are formed and tasks are processed in sequence through scheduling priority. The experimental results show that this strategy reduces the communication time and waiting time of running tasks in WSN. Meanwhile,the success rate of the task scheduler is improved and the efficiency of the system is optimized effectively.
出处
《传感技术学报》
CAS
CSCD
北大核心
2014年第3期378-382,共5页
Chinese Journal of Sensors and Actuators
基金
国家自然科学基金项目(61172058)
住房与城乡建设部研究开发项目(2009-K9-25)
关键词
任务调度
图灵机
有向无环图
智能建筑
WSN
WSN
task scheduling
turing machines
the directed acyclic graph
intelligent building
