移动云计算中能效感知的计算卸载机制研究

移动云计算可以将任务从移动设备计算卸载至云端以增强设备计算能力，而如何实现能效计算卸载机制是当前的主要挑战。为了解决该问题，以降低移动设备能耗和应用完成时间为目标，将计算卸载问题形式化为满足任务顺序与截止时间约束的能效代价最小化问题，并提出一种动态能效感知计算卸载算法。算法由三个子算法组成：计算卸载选择、时钟频率控制及传输功率分配。实验结果表明，通过局部计算时优化调整移动设备CPU时钟频率，以及云端计算时自适应分配传输功率，新算法可以有效降低应用执行能效代价，同时确保满足约束条件，提高执行效率。

Energy-aware computation offloading mechanism in mobile cloud computing

Mobile cloud computing can significantly enhance computation capability of mobile devices by offloading computation from the mobile devices onto the cloud. How to achieve energy-efficient computation offloading remains a challenge issue. For solving this problem, with reducing energy consumption and shorting application completion time as an objective, this algorithm was proposed. The algorithm will formulate computation offloading problem into the energy-eficiency cost minimization problem while satisfying the task-dependency requirements and the completion deadline constraint. And, a dynamic and energy-aware computation offloading algorithm is presented. The algorithm consists of three sub-algorithm of computation offloading selection, clock frequency control and transimission power allocation. Experimental results show that the algorithm can effective reduce the energy-eficiency cost by optimally adjusting the CPU clock frequency of mobile devices in local computing, and adapting the transmission power in cloud computing while ensures to satisfy the constraints and enhance the scheduling efficiency.