Resource Allocation in Multi-access Edge Computing: Optimization and Machine Learning

Abstract

Multi-access edge computing (MEC) equipped with artificial intelligence is a promising technology in B5G wireless systems. Some refined investigations and analysis are needed to gain more insights. This paper addresses that the core concept could be stemmed from the wait-and-see model in stochastic programming and indicates the quasi-separable property. Moreover, both small-scale fading and pathloss issues are included in the investigations. Two aspects of this study are the optimization model itself, followed by the simulation with machine learning (ML). One of the main interests of using ML is in improving the computational efficiency. Simulations showed that the efficiency may be improved from 93% to 96%.