Applying Model-Free Reinforcement Learning Algorithm in Network Slicing for 5G

A network slicing (NS) represents a form of isolated logical or virtual network architecture for a fixed physical network of 5G to support different radio access networks and adequately provide a specific service platform for the customers. The possible application of NS will find in the varying infrastructure of a network. Moreover, among several research challenges in NS, e.g., trouble to isolate the slices, mobility management, cooperation with other 5G technologies, and management of NS; the most challenging tasks are allocating resources of the slice efficiently and ensuring security. Reinforcement learning (RL) approach in NS for 5G helps encounter these critical challenges. It automatically reduces the latency and block suspected threat profile to ensure security in that slice. Additionally, the possible uses of allocated resources in SN improved sufficiently by applying RL technique. This paper will appropriately address to develop a policy-gradient based model-free RL approach to achieve optimal policy for NS with current values and action variables and accomplish optimal solutions for security improvement and detecting malicious nodes within the NS for 5G.