Computational Task Offloading Scheme based on Deep Learning for Financial Big Data

Mobile Edge Computing (MEC) technology reduces the computing delay of wireless devices by enabling devices offload resource-intensive financial big data computing tasks to the MEC server for execution, thereby greatly improving computing efficiency. This paper considers a wireless powered MEC system composed of an edge server and multiple wireless devices, and designs an optimal offloading strategy that assigns the computing mode of each wireless device (i.e., local computing or offloading to the edge server), so that the sum of the calculation rates of all wireless devices is the largest. Specifically, this paper first formalizes an optimization problem that maximizes the whole computation rates, and the decomposes the optimization problem into two sub-problems of optimal offloading decision and optimal time allocation. Thereafter, it proposes an offloading decision algorithm based on deep learning to solve the two sub-problems. In order to further reduce the computational complexity, this paper proposes an adaptive process that dynamically adjusts the algorithm parameters. The simulation results show that the proposed solution can quickly obtain offloading strategy and achieve load balancing. Compared with the existing related methods, this solution can significantly reduce the calculation time delay while obtaining near-optimal calculation performance.