Network Selection and Resource Allocation for Coexistence of eMBB and URLLC Services in a 6G Multi-Band HetNet

Abstract

In this paper, we study the energy-efficient joint network selection, sub-carrier allocation, and power control (EEJNSAPC) problem for the coexistence of enhanced mobile broadband (eMBB) and ultra-reliable low latency (URLLC) services in a shared 6G multi-band HetNet. To support such heterogeneous services, mobile network operators (MNOs) utilize a network sharing configuration known as the multi-operator core network that allows MNOs to share their radio access networks and spectrum resources. The goal is to ensure the minimum data rate requirement for eMBB users and guarantee the delay and reliability requirements for URLLC users. Also, leveraging stochastic geometry, a statistical quality of service analysis is provided for eMBB and URLLC users. Since the EEJNSAPC problem is a mixed-integer non-convex programming problem, it is challenging to be solved directly. Therefore, we propose a hybrid deep reinforcement learning (HDRL-EEJNSAPC) algorithm composed of a double dueling deep Q-network and a soft actor-critic to solve the EEJNSAPC problem. The simulation results reveal the performance of the HDRL-EEJNSAPC algorithm in terms of energy efficiency. Also, the superiority of HDRL-EEJNSAPC over reinforcement learning approaches is demonstrated via simulation results. Besides, simulation results confirm that HDRL-EEJNSAPC obtains performance close to optimal.