An Edge Computing Resource Offloading Method for Physical Layer Intelligent Communication Tasks

With the proliferation of smartphones and advancements in mobile communication networks, the cruciality of devising an efficient task offloading strategy for energy conservation and computational acceleration has been recognized. In a Mobile Edge Computing (MEC) network scenario consisting of a single mobile terminal and multiple micro base stations, an edge computing resource offloading method for Physical Layer smart communication tasks has been proposed. This approach considers the offloading ratio, decision, and channel allocation of User Equipment (UE), while leveraging the characteristics of Edge Computing (EC) to ensure Quality of Service (QoS) for users. To minimize system processing delay as the optimization objective, a Deep Reinforcement Learning (DRL) algorithm has been employed. Experimental simulations have shown that the proposed computational offloading method converges faster than existing algorithms, effectively reducing the total processing delay under similar conditions and achieving the lowest average delay value. Notably, with variations in local UE computational resources, the DRLO algorithm has demonstrated its ability to maintain the system's total delay at a relatively low level.