Computation Efficiency Maximization for Multi-RIS Assisted Integrated Communication and Computing Maritime IoT Networks

Reconfigurable intelligent surface (RIS) and edge computing have been envisioned as promising technologies for supporting integrated communication and computing maritime Internet of Things (ICC-MIoT) networks. In this paper, we propose an intelligent resource management scheme for multi-RIS assisted ICC-MIoT networks where each unmanned aerial vehicle (UAV) equipped with a RIS is deployed to set up reconfigurable links between unmanned surface vehicles (USVs) and the edge server due to the difficulties in establishing line-of-sight links. Then, an optimization problem is formulated to maximize the computation efficiency by jointly optimizing the association pairings, computation task assignment, computation resources, transmission powers, phase shifts, and trajectories of UAVs. To solve the formulated mixed-integer nonlinear programming problem involving both discrete and continuous decision variables, we propose a novel deep reinforcement learning algorithm based on deep Q network (DQN) algorithm and twin delayed deep deterministic (TD3) algorithm. Numerical results demonstrate the superiority of our proposed intelligent resource management scheme in enhancing the computation energy efficiency.