| Full text | |
| Author(s): |
Almeida, Gabriel M.
;
Camilo-Junior, Celso
;
Correa, Sand
;
Cardoso, Kleber
Total Authors: 4
|
| Document type: | Journal article |
| Source: | ICC 2023-IEEE INTERNATIONAL CONFERENCE ON COMMUNICATIONS; v. N/A, p. 6-pg., 2023-01-01. |
| Abstract | |
The virtualized radio access network (vRAN) placement problem can be defined as the joint decision of choosing the functional splits of the radio stack, where to run the virtualized functions of vRAN nodes, and the paths connecting the base stations with their respective protocol stacks. This optimization problem has been widely investigated in the literature with exact and heuristic approaches. While exact approaches still present very limited scalability, heuristic approaches achieve results still notably far from optimal. Metaheuristic techniques tend to be successful in this context, and an evolutionary approach has already shown promising results in a simplified version of the problem. In this work, we also employ a genetic algorithm to solve the vRAN placement problem but use a flexible formulation of the vRAN placement problem. We compare our proposal with two exact approaches and one heuristic approach (based on machine learning) from the literature. Our proposal is able to solve large instances of the problem in a reasonable time while achieving satisfactory results, close to the optimal. Additionally, with our knowledge of the problem, we created synthetically a single individual in the first generation which made it possible to obtain a high-quality (i.e., close to the optimal) first solution for several instances. (AU) | |
| FAPESP's process: | 20/05127-2 - SAMURAI: smart 5G core and multiran integration |
| Grantee: | Aldebaro Barreto da Rocha Klautau Junior |
| Support Opportunities: | Research Projects - Thematic Grants |