An integrated approach for dual resource optimization of relay‐based mobile edge computing system

Summary The evolution of IoT, 5G and 6G aims to provide almost zero latency. Computation tasks size is different for different users. A framework for task computation achieving almost zero latency for stochastic demand is a challenge. A relay‐based D2D mobile edge computing (MEC) system is proposed. The idle device present in the networks is used as relay resources (RS). Mobile devices (MDs) communicate task to relay resources (RS) using D2D communication link. The RS perform computation and offloaded to edge server (ES). It aims to minimize total cost, energy expenditure and overall latency. Problem is formulated as mixed‐integer nonlinear‐constrained problem (MINCP). A three‐step algorithm to optimize relay selection, power allocation and computation resource allocation is proposed. In the initial step optimal relay selection is obtained by the Kuhn‐Munkres (KM) algorithm. In the next step, power allocation is obtained using Q‐learning. In the last step, the main problem is converted into a cost optimization problem deciphered by the proposed algorithm. The substantial simulation results indicate the relay‐based MEC system to achieve an astounding outcome in terms of latency, energy consumption and cost. Compared with other baseline methods, the proposed algorithm can achieve reduced energy consumption and cost for almost zero latency.