Dynamic Resource Management in Next Generation Networks with Dense User Traffic

Abstract

With the era of the fifth generation (5G) networks, supporting all mobile service users who have different Quality of Service (QoS) requirements becomes the main challenge. To manage and satisfy the heterogeneous requirements, network slicing concept can be a solution over a common physical infrastructure. Splitting the network into slices which have different properties (e.g., bandwidth requirements, delay tolerance, user density, etc.) allows to schedule and optimize the requests under the constraint of limited resources. The network has to decide to accept or reject the requests, and scale up/down the slices by considering the user density in accepted requests, and then, schedule the accepted requests to serve them in an order. In this paper, it is verified that slicing the network and scaling up/down the slices by using deep reinforcement learning (DRL) algorithms with consideration of user density, improve the speed of satisfaction of users with respect to the classical baseline scheduling algorithms.