嵌入式异构环境下任务调度算法的研究

A Task Scheduling Algorithm in Embedded Heterogeneous Environments

高效的任务调度是提升系统性能的关键因素之一。讨论在任务异构和通信速度差异的Fork-join型嵌入式环境下,独立任务的调度问题,提出新的分配调度方案,选取负载最小的处理节点进行分配,实现节点间的负载均衡,且满足任务的响应时间和处理节点数目最小化的要求。基于方案,构造一个以任务的平均响应时间驱动的启发式算法:ARTDHA(Aver-age-Response-Time-Driven Heuristic Algorithm)。仿真实验表明,算法更符合复杂的嵌入式异构环境,能更好满足系统的时间特性、最小化资源的开销,同时任务的调度时间要优于FCFS(First Come First Serve)算法。

Efficient task scheduling is one of the key issues for achieving high performance of system. Independent task scheduling in fork -join -based heterogeneous environments with diverse tasks and different communication speed is discussed and a new scheduling scheme is presented. It selects the node with the minimum load to distribute the task, thus achieving the load balancing, meeting the demand of the response time and minimizing the number of nodes. Based on the scheme, the paper presents an Average - Response - Time - Driven Heuristic Algorithm- ARTDHA. Experimental results show that in the embedded heterogeneous environment ARTDHA algorithm performs much better than FCFS （First Come First Serve） algorithm in the aspect of scheduling time.