基于马尔可夫决策的动态电源管理技术

Dynamic Power Management Based on Markov Process

  • 摘要: 研究了随机最优算法,并建立了该算法的电源管理系统模型。在系统任务响应延迟的约束条件下,应用马尔可夫决策方法对电源管理方案进行优化,得出系统模型最佳的电源管理策略。分析比较表明,一些传统的动态电源管理方法虽然也可以有效地减少系统的能耗,但是对系统的性能却有较大的影响。随机最优算法可以在节约系统能耗和保证系统性能之间取得较好的平衡。

     

    Abstract: The stochastic optimization algorithm is studied and the model of the power management system is presented. IN the situation of system's response-delay, Markov decision method is applied to optimize the power management plan and the best power management strategy is educed. The analysis of the comparison demonstrates that despite efficient energy reduction in some dynamic power management there is considerable negative influence on the system's performance. The stochastic optimization algorithm can get favorable equilibrium between the energy consumption and the reliability of performance in the system.

     

/

返回文章
返回