Multi-Objective Multi-Dimensional Resource Allocation for Categorized QoS Provisioning in Beyond 5G and 6G Radio Access Networks

• Published in: IEEE transactions on communications Vol. 72; no. 3; pp. 1790 - 1803
• Main Authors: Fu, Yongqin, Wang, Xianbin, Fang, Fang
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.03.2024; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Backhaul networks; Categories; categorized QoS provisioning; Computation; Data communication; Decoupling; decoupling-based iterative optimization; Downlink; Heuristic methods; Iterative methods; Linear programming; Mixed integer; mixed-integer non-linear programming; Multi-objective multi-dimensional resource allocation; Multiple objective analysis; Nonlinear programming; Optimization; Provisioning; Quality of service; Quality of service architectures; radio access network; Resource allocation; Resource management; Ultra reliable low latency communication; Virtual networks
• ISSN: 0090-6778; 1558-0857
• DOI: 10.1109/TCOMM.2023.3335414

To effectively meet the diverse Quality of Service (QoS) requirements from proliferating applications, a widely-adopted practical solution in radio access network (RAN) is categorized QoS provisioning, which utilizes virtual networks (i.e., tenants) to support a limited number of service categories. Apparently, one critical issue is RAN resource allocation among coexisting tenants. However, conventional single objective-based approaches cannot ensure fairness among different service categories. Moreover, except for radio resource, computing and storage resources also need to be considered. Besides, appropriate allocation of computing and storage resources could help mitigating backhaul network congestion. Hence, we aim to optimize the key QoS indicators of three main service categories and reduce backhaul bandwidth consumption simultaneously. We formulate the problem of multi-dimensional resource allocation from RAN to tenants as a multi-objective mixed-integer non-linear programming (MINLP) problem, which is challenging to solve directly due to the competing objectives and the mutual-influenced resources. For guaranteeing fairness, this problem is reformulated as a single-objective optimization problem using weighted sum approach. Moreover, a decoupling-based iterative optimization (DBIO) algorithm is proposed to decompose it into three subproblems to solve iteratively. Simulation results demonstrate that DBIO algorithm can achieve superior performance with much less time consumption, compared with three metaheuristic algorithms.