• 论文 •    

一类资源负荷均衡问题的优化调度模型及其算法

曹华军,刘飞,施金良   

  1. 1.重庆大学 制造工程研究所,重庆400030;2.重庆科技学院 电子信息工程学院,重庆400050
  • 出版日期:2005-05-15 发布日期:2005-05-25

Optimization scheduling model and its algorithm for a kind of resource workload balancing problems Optimization scheduling model and its algorithm for a kind of resource workload balancing problems Optimization scheduling model and its algorithm for a kind of resource workload balancing problems

CAO Hua-jun, LIU Fei, SHI Jin-liang CAO Hua-jun, LIU Fei, SHI Jin-liang CAO Hua-jun, LIU Fei, SHI Jin-liang   

  1. 1. Inst. of Manu. Eng., Chongqing Univ., Chongqing400030, China;2. Coll. of Electronics Info. Eng.,Chongqing S & T Univ.,Chongqing400050, China
  • Online:2005-05-15 Published:2005-05-25

摘要: 针对多个独立任务在多个不完全同等的处理机上处理时,处理机的最大负荷为最小的非抢先调度问题,建立了一类资源负荷均衡问题的优化调度模型。该模型引入0-1方案矩阵和时间负荷矩阵,分别描述了独立任务分配问题和负荷调度问题;针对部分处理机不能处理某一个独立任务的情形,假定其单位处理时间负荷趋于无穷大,从而避免优化调度中出现伪解。采用遗传算法对模型进行了求解。为了提高遗传算法的运算效率,采用整数方案描述和二进制间接编码的方法对方案染色体进行编码。最后,通过一个案例对模型和算法的有效性进行了验证。

关键词: 负荷均衡, 调度模型, 遗传算法

Abstract: A novel optimization scheduling model and its algorithm were developed for a kind of resource workload balancing problems, in which many independent tasks should be scheduled non-preemptively on non-identical multiprocessors to minimize the maximum workload. This model introduced a 0-1 scheme matrix and a workload matrix to describe task assignment and time load scheduling respectively. Under the circumstance that some processors can not process some specific tasks,infinite time load values were assumed to avoid false solutions. Genetic algorithm (GA) was applied to solve the scheduling model. To improve the efficiency of GA, the scheduling scheme chromosome was coded indirectly, in which the decimal digits were used to describe the assignment relationship of task sand processors, and the binary digits were used to code the decimal digits. Finally, a case study was presented to validate the model and its algorithm.

Key words: workload balancing, scheduling model, genetic algorithm

中图分类号: