Joint Task Allocation and Computation Offloading in Mobile Edge Computing With Energy Harvesting

Mobile edge computing (MEC) is a burgeoning paradigm that MEC servers provide the computing capabilities to release the workload of the mobile devices by transferring the computational tasks, which can vastly reduce the latency and energy cost for executing tasks. In consideration of the battery capacity limitation with the mobile devices, the computation task process may be interrupted. To improve the computational service capacity as well as the popularity of the green computing, the energy of mobile devices is considered to be supplied effectively by energy harvesting (EH), capturing the energy from the environment. We propose an effective task allocation strategy that minimizes the weight sum of energy cost and computational latency of mobile devices in an MEC system with EH. Furthermore, we construct a task queue to fetch the upcoming tasks for mobile devices. On the basis of the Lyapunov optimization approach, we propose an online Lyapunov optimization-based dynamic task allocation (LODTA) algorithm that determines the task assignment policy through adjusting mobile devices with the CPU execution frequency and the transmission power caused by offloading. The LODTA algorithm has a superiority that only the current system state is necessary for the task allocation strategy, but without predicting the future state. In our simulation, the proposed model and algorithm can stabilize the battery energy level with a trade-off between energy consumption and execution latency.