Machine Learning-Based Uplink Scheduling Approaches for Mixed Traffic in Cellular Systems

The problem of the uplink resource allocation mixed-traffic types in cellular networks is a challenging that has not been addressed sufficiently literature. In this paper, we consider 5G scheduling for Ultra-Reliable and Low Latency Communications enhanced Mobile Broad-Band (eMBB) traffic types. There are three main techniques to be considered, namely, grant-based (GB), semi-persistent, grant-free (GF) techniques. Furthermore, there different schemes used GF scheduling, reactive, k-repetitions, proactive schemes. We devise mathematical model services using k-repetitions scheme as first define such single cell. addition, GB eMBB adapted fit our problem. formulate mixed-integer non-linear programming optimization introduce complete system includes subsystems. novel mixed scheduler combines advantages two well-known schedulers machine-learning-based algorithms evaluate them comparison literature addition optimal bound also derive. results show proposed produce near-optimal real time.