调度In-Tree图的启发函数构造

Constructing a heuristic function to schedule In_Tree graph

在分析了其它同类启发函数的构造机理后,克服"以当前任务具有最早起始执行时刻"贪心方法构造启发函数的缺点,提出了基于任务复制,"以当前任务的直接后继具有最早起始执行时刻"为目标来构造动态启发函数的方案.通过理论证明与模拟实验表明:所构造的启发函数具有较强的启发能力,对In＿Tree图的调度具有较好的效果.这对研究实时分布式计算机系统的任务调度算法具有理论和实际意义.

Analyzes the mechanism of some typical heuristic functions, and puts forward the duplication-based method to construct a dynamic heuristic function which lets the immediate successor of the current task have the earliest start time,overcoming the flaw of greed method that the current task has the earliest start time. Proven by simulated experiment, the constructed heuristic function has strong capacity, especially for scheduling In_Tree graph. This will help schedule the tasks of real-time distributed system theoretically and practically.