##Admission Control and resource Allocation for 5G Core Network Slicing based on Deep Reinforcement Learning

Network Slicing is a promising technology forproviding customized logical and virtualized networks for theindustry’s vertical segments.
Distinct use cases in 5G networks, such as enhanced Mobile Broadband, Ultra-Reliable Low LatencyCommunications, and Massive Internet of Things, have theirQuality of Service requirements, which must be supported. This approach encompasses mechanisms for Admission Control and Resource Allocation fornetwork slicing in 5G core networks.

The Admission Control mechanism introduces two solutions for learning the admissionpolicy that optimizes the profit of Network Slice Providers, one based on Reinforcement Learning (called SARA) and the otherbased on Deep Reinforcement Learning (called DSARA). The SARA and DSARA solutions consider the 5G use cases defined bythe 3rd Generation Partnership Project. The Resource Allocationmechanism tries to balance the load on the network nodes as wellas minimize the network resource utilization.