Deep reinforcement learning for resource allocation with network slicing in cognitive radio network

Abstract

With the development of wireless communication technology, the requirement for data rate is growing rapidly. Mobile communication system faces the problem of shortage of spectrum resources. Cognitive radio technology allows secondary users to use the frequencies authorized to the primary user with the permission of the primary user, which can effectively improve the utilization of spectrum resources. In this article, we establish a cognitive network model based on under1 lay model and propose a cognitive network resource allocation algorithm based on DDQN (Double Deep Q Network). The algorithm jointly optimizes the spectrum efficiency of the cognitive network and QoE (Quality of Experience) of cognitive users through channel selection and power control of the cognitive users. Simulation results show that proposed algorithm can effectively improve the spectral efficiency and QoE. Compared with Q-learning and DQN, this algorithm can converge faster and obtain higher spectral efficiency and QoE. The algorithm shows a more stable and efficient performance.

The Frontiers of Deep Reinforcement Learning for Resource Management in Future Wireless HetNets: Techniques, Challenges, and Research Directions

“…Another pioneering work is reported in [181] in which the authors propose a single-agent Double DQN-based DRL to address the problem of joint channel selection and power allocation with network slicing in CRNs. Their study aims to provide high SE and QoS for cognitive users.…”

Applications of Machine Learning in Resource Management for RAN-Slicing in 5G and Beyond Networks: A Survey

“…The proposals allocate the radio and power resources to the slices. In [112], a method called CNDDQN based on DDQN (Double Deep Q Network) is proposed for resource allocation in cognitive RAN slicing. In this method, radio and power resources are allocated to eMBB and uRLLC slices by considering SE and QoE of users (which is defined as the rate of sent packets to the total number of packets for each user).…”

“…In RL methods, GAN-DQN ( [127]) and GAT-DQN ( [91], [92]) methods have been used to improve decision making in DQN methods. Also, Double DQN ( [115], [112], [84]) and Dueling DQN ( [106], [113], [93], [135]) have also been used to improve decision-making and overestimation in DQN-based methods. Due to the use of AC methods in [79] - [81], decisions may fluctuate, so A2C and DDPG are used to address this problem in [110], [82], [115], [116], [83], [107], [132].…”

“…In recent years, the use of distributed methods (such as transfer learning and federated learning) in ML has been considered due to the common challenges in centralized methods [79], [80], [81], [82], [84], [85], [86], [87], [88], [89], [90], [91], [92], [83], [93], [94], [95], [110], [111], [112], [113], [114], [115], [116], [117], [118], [136] ✓ ✓ ---- [96], [97], [98], [99], [100], [103], [104], [105], [123], [124], [135] [107], [108], [109], [126], [127], [128], [35], [129], [130] such as centralization, security, privacy, and time and computational complexity. In distributed methods, a ML mode...…”

Deep Reinforcement Learning for Radio Resource Allocation and Management in Next Generation Heterogeneous Wireless Networks: A Survey

“…Using experimental results, it is shown that agents learn to satisfy the latency constraints on V2V links while minimizing the interference to V2I communications. The authors in [176] propose a single-agent Double DQNbased DRL to address the problem of joint channel selection and power allocation with network slicing in CRNs. The aim of their study is to provide high SE and QoS for cognitive users.…”