Digital Twin Empowered Task Offloading for RIS-Assisted Edge Computing Networks

In order to meet the high computing demands caused by emerging compute-intensive applications in Mobile Edge Computing (MEC), this paper proposes a Digital Twin (DT)-empowered task offloading scheme where Reconfigurable Intelligent Surface (RIS) is used to enhance the communication links and extend the coverage. Firstly, the joint optimization of user offloading strategy, RIS phase-shift vector, beamforming vector, transmit power of users and computation capacity allocation are investigated with the aim of minimizing the total energy consumption of users and resource devices under the constraints of communication and computing resources. Then, the formulated non-convex combinational optimization problem is decomposed into three sub-problems, including RIS phase-shift design, binary optimization of transmit power, and computing resource allocation. In addition, the Double Deep Q Network (DDQN) approach is invoked to determine the offloading decisions and an alternating iteration optimization algorithm is designed to achieve the optimal solution. Simulation results show that the DDQN-based algorithm is able to train quickly and reduce effectively the total energy consumption of the system.