动态电源管理预测算法设计

摘要

Efficient dynamic power management and prediction algorithm has a great improvement on the life time of the equipment,and it is the key of dynamic power management.Under this background,a self feedback embedded Markov algorithm,which was a dynamic power management predictive algorithm and based on the characteristics of multi task system,was proposed.Its running status was classified.Abstract data model and data model were constructed by Markov algorithm.Based on the constructed model,the transition probability matrix with the function of using the initial state to predict the future state function was obtained.And the self feedback mechanism was added,which made the algorithm's robustness and prediction accuracy significantly improved compared with before.%高效的动态电源管理预测算法对设备的续航时间有极大的提升,是动态电源管理研究的关键.在此背景下,提出了自反馈内嵌式马尔可夫(Markov)算法,它是一种动态电源管理的预测式算法.该算法针对多任务系统的特性,对其运行状态进行归类.抽象数据模型和实例数据模型通过Markov算法构建,在构建模型的基础上,得到具有利用初始状态预测未来运行状态功能的转换概率矩阵.并加入自反馈机制,使得与改进前相比,算法的健壮性和预测精度都有显著的提升.