一个启发式任务团调度算法

A heuristic task-clustering scheduling algorithm

基于任务聚类的启发式任务团调度是实现并行任务调度的有效方法。面向大型复杂武器系统的实时并行仿真，在研制自动并行化系统AFP的过程中，基于关键任务的聚类算法，设计并实现了一个启发式任务团调度算法。该算法以启发式调度函数最小为目标，综合考虑了各任务团的计算开销和相互之间的通信开销，降低了调度算法的复杂性和NPC问题而带来的回溯性，并可以获得满意的并行效率。

Using heuristic task-clustering scheduling is an efficient method for parallel tasks scheduling.Based upon real-time parallel simulation of large-scale and complex weapon system, in the course of developing the automatic paralyzing system AFP, we propound a heuristic taskclustering scheduling algorithm.By considering computation cost and commnicaton cost synthtically it aims the minimum value of the heuristic scheduling funtion, can reduce complexity of scheduling problem and acquire satisfied parallel efficiency This paper gives its main idea and exploring results.