Deep Reinforcement Learning for Adaptive Network Slicing in 5G for Intelligent Vehicular Systems and Smart Cities

Intelligent vehicular systems and smart city applications are the fastest growing Internet-of-Things (IoT) implementations at a compound annual growth rate of 30%. In view recent advances in IoT devices emerging new breed driven by artificial intelligence (AI), fog radio access network (F-RAN) has been recently introduced for fifth-generation (5G) wireless communications to overcome latency limitations cloud-RAN (C-RAN). We consider slicing problem allocating limited resources edge (fog nodes) users with heterogeneous computing demands dynamic environments. develop model based on cluster nodes (FNs) coordinated an controller (EC) efficiently utilize edge. For each service request cluster, EC decides which FN execute task, i.e., locally serve edge, or reject task refer it cloud. formulate as infinite-horizon Markov decision process (MDP) propose deep reinforcement learning (DRL) solution adaptively learn optimal policy. The performance proposed DRL-based method is evaluated comparing other approaches environments different scenarios design objectives. Comprehensive simulation results corroborate that quickly learns policy through interaction environment, enables adaptive automated efficient resource allocation