AI-Based Resource Allocation in End-to-End Network Slicing Under Demand and CSI Uncertainties

Amir, Gharehgoli,; Nouruzi, Ali; Mokari, Nader; Azmi, Paeiz; Javan, Mohammad Reza; Jorswieck, Eduard A.

doi:10.1109/tnsm.2023.3243837

Cited by 11 publications

(9 citation statements)

References 72 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The proposed framework in [31] aims to maximize the resource allocation efficiency through slices and the user's quality of experience. In [32], an RL-based framework is proposed focusing mainly on eMBB aiming to solve resource allocation problem that are defined with Channel State Information (CSI) uncertainty. The authors model the CSI uncertainty by three methods: worst-case, probabilistic, and hybrid.…”

Section: Literature Reviewmentioning

confidence: 99%

Joint Q-Learning Based Resource Allocation and Multi-Numerology B5G Network Slicing Exploiting LWA Technology

Elmosilhy,

Elmesalawy,

Ibrahim

et al. 2024

IEEE Access

View full text Add to dashboard Cite

The emergence of the sixth generation (6G) era has highlighted the importance of Network Slicing (NS) technology as a promising solution for catering the diverse service requests of users . With the presence of a large number of devices with different service requests and since each service has different goal and requirements; efficiently allocating Resource Blocks (RBs) to each network slice is a challenging task in order to meet the desired Quality of Service (QoS) standards. However, it is worth noting that the majority of research efforts have primarily concentrated on cellular technologies, leaving behind the potential benefits of utilizing unlicensed bands to alleviate traffic congestion and enhance the capacity of existing LTE networks. In this paper we propose a novel idea by exploiting LTE-WLAN Aggregation technology (LWA) in Multi-Radio Access Technology (RAT) Heterogeneous Networks (HetNet), aiming to solve radio resource allocation problem based on the Radio Access Network (RAN) slicing and 5G New Radio (NR) scalable numerology technique. A joint optimization problem is proposed by jointly finding an efficient resource allocation ratio for each slice in each Base Station (BS) and by finding the optimum value of scalable numerology with the objective of maximizing users' satisfaction. In order to solve this problem, a novel three-stage framework is proposed which is based on channel state information as a pre-association stage, Reinforcement Learning (RL) algorithm as finding the optimum value of slice resource ratio and scalable numerology, and finally Regret Learning Algorithm (RLA) as users' re-association phase. Furthermore, a comprehensive performance evaluation is conducted against different base-line approaches. The simulation results show that our proposed model balance and achieve improvement in users' satisfaction by deploying the proposed Multi-RAT Het-Net architecture that leverage LWA technology.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Joint Q-Learning Based Resource Allocation and Multi-Numerology B5G Network Slicing Exploiting LWA Technology

Elmosilhy,

Elmesalawy,

Ibrahim

et al. 2024

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Then, we measured the MARDPG optimality gap using the result of the exhaustive search algorithm as a reference. Finally, in the last scenario, by calculating the fairness score, we can realize how much resources are fairly distributed among the users [73].…”

Section: Simulationmentioning

confidence: 99%

Resilient Disaster Relief in Industrial IoT: UAV Trajectory Design and Resource Allocation in 6G Non-Terrestrial Networks

Mohammadisarab,

Nouruzi,

Khalili

et al. 2024

IEEE Open J. Commun. Soc.

Self Cite

View full text Add to dashboard Cite

This research explores strategies to augment the connectivity among users within Hierarchical Non-Terrestrial Networks (HNTNs) dedicated to Disaster Relief Services (DRS). The primary goal is to optimize radio resources, computing capacities, and the trajectories of Unmanned Aerial Vehicles (UAVs) at each time slot, aiming to maximize the number of satisfactory connections (NSCs). UAVs function as aerial base stations (ABSs), establishing links for reduced capability (RedCap) user equipment (UE) using power domain non-orthogonal multiple access (PD-NOMA). Given the potential inoperability of terrestrial networks during disasters, swift data transmission is critical for mission-critical (MC) UEs. Therefore, end-to-end (E2E) delay is a crucial quality of service (QoS) constraint. The proposed problem is solved using a multi-agent recurrent deterministic policy gradient (MARDPG) algorithm, where the ABSs collaborate to maximize the NSCs and determine their optimal policy by interacting with the environment. Additionally, a sharing experience module (SEM) is incorporated, enabling agents to encode actions and observations using long short-term memory (LSTM), allowing each agent to utilize the historical actions and observations of other agents. To demonstrate the superiority of MARDPG, three algorithmic benchmarks and four different system models are implemented. The numerical results demonstrate the impact of various parameters, such as the number of subcarriers, users, and the maximum tolerable E2E delay on the NSCs. Furthermore, different scenarios indicate that MARDPG outperforms the benchmarks, achieving approximately a 6 percent optimality gap and a 91 percent fairness for achievable rate among users. HNTNs, DRS, UAVs, PD-NOMA, MARDPG I. INTRODUCTION A. Motivations and State-of-the-Art W ireless communication technologies have undergone significant advancements in recent years, emerging as a critical facilitator for upcoming consumer applications and services [2], [3]. The transition from the fourth-generation (4G) to the fifth generation leads to a categorizing of quality of services (QoSs), which are named ultra-reliable low-latency communications (URLLC), massive machine-type communication (mMTC), and enhanced mobile broadband

show abstract

“…Most studies on network slicing have investigated the key management functionalities of admission control of slice requests [10,18] and allocation of resources to individual slices [11,35] in isolation. Previous works that jointly addressed the two tasks [4,7,8] have overlook the important trade-offs entailed by: (𝑖) the added revenues of accepting slices that request capacity beyond that available, while not using it all the time; and, (𝑖𝑖) the potential cost of violating SLAs in the moments when the actual demand of all accepted slices exceeds the total capacity.…”

Section: Related Workmentioning

confidence: 99%

“…Discrepancies between (8) and the objective functions of (P1) and (P2) are possible due to slice traffic prediction errors, and are part of the complexity of an overbooking-based NSaaS.…”

Section: Mno Profit From Ac/ra Decisionsmentioning

confidence: 99%

kaNSaaS: Combining Deep Learning and Optimization for Practical Overbooking of Network Slices

Alcalá-Marín,

Bazco-Nogueras,

Banchs

et al. 2023

Proceedings of the Twenty-Fourth International Symposium on Theory, Algorithmic Foundations, and Protocol Design for Mobile Net

View full text Add to dashboard Cite

Cloud-native mobile networks pave the road for Network Slicing as a Service (NSaaS), where slice overbooking is a promising management strategy to maximize the revenues from admitted slices by exploiting the fact they are unlikely to fully utilize their reserved resources concurrently. While seminal works have shown the potential of overbooking for NSaaS in simplistic cases, its realization is challenging in practical scenarios with realistic slice demands, where its actual performance remains to be tested. In this paper, we propose kaNSaaS, a complete solution for NSaaS management with slice overbooking that combines deep learning and classical optimization to jointly solve the key tasks of admission control and resource allocation. Experiments with large-scale measurement data of actual tenant demands show that kaNSaaS increases the network operator profits by 300% with respect to NSaaS management strategies that do not employ overbooking, while outperforming by more than 20% state-of-the-art overbooking-based approaches.

show abstract

AI-Based Resource Allocation in End-to-End Network Slicing Under Demand and CSI Uncertainties

Cited by 11 publications

References 72 publications

Joint Q-Learning Based Resource Allocation and Multi-Numerology B5G Network Slicing Exploiting LWA Technology

Joint Q-Learning Based Resource Allocation and Multi-Numerology B5G Network Slicing Exploiting LWA Technology

Resilient Disaster Relief in Industrial IoT: UAV Trajectory Design and Resource Allocation in 6G Non-Terrestrial Networks

kaNSaaS: Combining Deep Learning and Optimization for Practical Overbooking of Network Slices

Contact Info

Product

Resources

About