Data Security Aware and Effective Task Offloading Strategy in Mobile Edge Computing

With the research and development of 5G technology, emerging markets such as Wise Information Technology of med, smart transportation and industrial Internet are gradually growing, which not only provide convenience to people’s life, but also put forward increasingly urgent demand for efficient parallel and distributed technologies. Therefore, in order to meet the need of high computing amount for application diversification, this paper proposes a novel scheduling solution with data security, aiming at simultaneously optimizing the system response time and the user’s energy consumption. First, we model the scheduling problem in a mobile edge computing (MEC) environment as a Markov decision process (MDP) problem, and a three-tier collaboration model considering data security in the MEC environment is constructed. Second, the system response time and the energy consumption are simultaneously optimized in this paper, with objective weights which change in real-time. At the same time, load balancing at the edge layer is considered. Third, a deep reinforcement learning (DRL)-based secure offloading (DRLSO) algorithm is given as the solution for the research problem. In experiments from multiple angles, the proposed algorithm has good performance.