Partial Critical Path Based Greedy Offloading in Small Cell Cloud

With mobile applications sharply developing, the battery technology becomes the bottleneck. Meanwhile, mobile users are increasingly sensitive to the latency of an application. The computation offloading in Small Cell Cloud (SCC) can economize the energy consumption of mobile devices efficiently and guarantee the makespan of an application. In this paper, we model the mobile application as a directed acyclic graph (DAG), and formulate an optimization problem of collaborative task execution to minimize the energy consumption on the mobile device while meeting a prescribed latency constraint. In order to solve this NP-hard problem, we propose a greedy algorithm based on partial critical path (GA-PCP) which can solve the problem approximately. The algorithm partitions the DAG into chains and processes these chains with the ``Add- Compare-Select" strategy to obtain the execution strategy. The algorithm can obtain a polynomial time complexity. Simulation results show that the solution of the GA-PCP is close to the optimal solution of the enumeration algorithm. Besides, the GA-PCP execution strategy can significantly save the energy consumption on the mobile device thereby prolonging its battery life, compared to the local execution.