Game-theoretic Learning-based QoS Satisfaction in Autonomous Mobile Edge Computing

Mobile Edge Computing (MEC) has arisen as an effective computation paradigm to deal with the advanced application requirements in Internet of Things (IoT). In this paper, we treat the joint problem of autonomous MEC servers' operation and mobile devices' QoS satisfaction in a fully distributed IoT network. The autonomous MEC servers' activation is formulated as a minority game and through a distributed learning algorithm each server determines whether it becomes active or not. The mobile devices acting as stochastic learning automata select in a fully distributed manner an active server to get associated with for computation offloading, while for energy efficiency considerations, a non-cooperative game of satisfaction form among the IoT devices is formulated to determine the transmission power of each device in order to guarantee its QoS satisfaction. The performance evaluation of the proposed framework is achieved via modeling and simulation and detailed numerical and comparative results demonstrate its effectiveness, scalability, and robustness.