基于马尔科夫链的构件化嵌入式软件能耗估算模型  

Component-based Embedded Software Energy Estimated Model Based on Markov Chain

作　　者：熊冰[1] 郭兵[1] 沈艳[2] 王继禾[1] 伍元胜[1] 刘云本[2] 

机构地区：[1]四川大学计算机学院,成都610065 [2]成都信息工程学院控制工程学院,成都610225

出　　处：《小型微型计算机系统》2012年第3期655-659,共5页Journal of Chinese Computer Systems

基　　金：国家"八六三"高技术研究发展计划项目(2008AA01Z105)资助;国家自然科(61073045)资助;四川省杰出青年科技基金项目(2010JQ0011)资助

摘　　要：在目前全球倡导"低碳经济"的背景下,嵌入式软件能耗已成为嵌入式系统设计的重要考量因素,一种快速有效的软件能耗估算模型对于嵌入式软件早期开发具有重要的意义.本文着眼于构件化嵌入式软件,从基于状态的角度出发,将嵌入式软件系统的运行过程视为一个状态的集合,提出了一种基于马尔科夫链的嵌入式软件能耗估算模型.然后,通过使用状态转移概率矩阵获得嵌入式软件构件的稳定状态分布,再结合构件的能耗测量值,实现嵌入式软件的能耗估算,并通过实验验证了该模型的有效性.Nowadays as low carbon economy has been advocated worldwide,the energy consumption of embedded software has become a crucial factor in embedded system design.A quick and efficient energy estimation model is important to the early development of embedded software.Firstly,Oriented to component-based embedded software,by considering the running procedure of embedded software as a state set,this paper proposes a Markov-Chain-based embedded system energy estimation model.Then,integrating the energy consumption measuring values of components and steady state distribution of embedded software obtained by using a state transition probability matrix,this model achieves the goal of energy consumption estimation.Lastly,the experimental results show that this model is effective.

关 键 词：嵌入式软件能耗 构件化嵌入式软件 能耗估算 马尔科夫链 

分 类 号：TP302.7[自动化与计算机技术—计算机系统结构]