Machine learning-based algorithm for core allocation in spatial division multiplexing elastic optical networks
Júnior, J.
; Sousa, C.
; Morais, A.
;
Cartaxo, A.
; Soares, A.
Optical Fiber Technology Vol. 91, Nº , pp. 104155 - 104155, May, 2025.
ISSN (print): 1068-5200
ISSN (online):
Scimago Journal Ranking: 0,54 (in 2025)
Digital Object Identifier: 10.1016/j.yofte.2025.104155
Abstract
Spatial division multiplexing elastic optical networks (SDM-EONs) using multicore fibers (MCF) are promising candidates for the future transport networks. In MCFs, a new dimension is added to the resource allocation problem: core allocation. In this paper, a machine learning-based algorithm for core selection (MaLAC) in SDM-EONs is proposed. Compared with other three solutions proposed in the literature and a scenario with a low crosstalk level, MaLAC achieves at least 25.35% gain in terms of request blocking probability (RBP) and at least 24.81% for bandwidth blocking probability (BBP). In a scenario with a high crosstalk level, MaLAC achieves at least 8.16% gain for RBP and at least 9.28% for BBP.
