- Advanced Optical Network Technologies
- Optical Network Technologies
- Advanced Photonic Communication Systems
- Software-Defined Networks and 5G
- Semiconductor Lasers and Optical Devices
- Adversarial Robustness in Machine Learning
- Explainable Artificial Intelligence (XAI)
- Data Visualization and Analytics
- Islanding Detection in Power Systems
Federico Santa María Technical University
2020-2024
University of Valparaíso
2020-2024
Universidad Técnica Federico Santa María
2022
Multi-band elastic optical networks are a promising alternative to meet the bandwidth demand of ever-growing Internet traffic. In this letter, we propose family band allocation algorithms for multi-band networks. Employing simulation, evaluate blocking performance 3 such and compare their with only heuristic proposed date. Results show that three new outperform previous proposal, up one order magnitude improvement. We expect these results help advance area dynamic resource in
The use of additional fiber bands for optical communications -known as Multi-band or Band-division multiplexing (BDM) - allows to increase the traffic served in transparent networks. In recent years, many proposals have emerged a solution resource allocation such multi-band architectures. This work presents novel approach based on reinforcement learning (RL) techniques accommodate elastic network resources. Two new environments were implemented and added Optical-RL-Gym toolkit considering...
This paper studies the impact of demand-prioritization on Space-Division Multiplexing Elastic Optical Networks (SDM-EON). For this purpose, we solve static Routing, Modulation Level, Spatial Mode, and Spectrum Assignment (RMLSSA) problem using 34 different explainable strategies. Although previous works have applied heuristics or meta-heuristics to perform demand-prioritization, they not focused identifying best prioritization strategies, their inner operation, implications behind good...
Deep Reinforcement Learning (DRL) has proven a considerable potential for enabling non-trivial solutions to resource allocation problems in optical networks. However, applying plain DRL does not ensure better performance than currently known best heuristics solutions. demands parameter tuning process improve its performance. One possibility is the reward function design. The allows feedback agents on whether actions sent environment were successful or not. A transparent returns action...
Space division multiplexing (SDM) systems using few-mode fibers (FMF) are essential for next-generation fiber optic communications. Optical amplifiers with low noise, minimal differential modal gain (DMG), and spectral (DSG) these systems. In this work, we present a method to design optimize two-stage erbium-doped amplifier (FM-EDFA) joint DMG-DSG minimization approach. This methodology involves the pumping profile flattening filter design. Simulation results show FM-EDFA achieves DMG DSG...
A highly accurate analytical procedure to support emerging ZR+-enabled IPoWDM network planning tools is presented. The four orders of magnitude faster than simulation, allowing extensive and fast analysis design.
A highly accurate analytical procedure to support emerging ZR+-enabled IPoWDM network planning tools is presented. The four orders of magnitude faster than simulation, allowing extensive and fast analysis design.
In this paper, we analyze the impact of sorting network users when attending their demands on performance. For purpose, solve Routing, Modulation Level, Spatial mode, and Spectrum Assignment (RMLSSA) problem for elastic optical networks (EONs) with spatial-division multiplexing (SDM) static operation. We use several different criteria to order find best ones in terms capacity spectrum spatial fragmentation network. Experimental results that a descending obtains better results,...
Network virtualization is a key enabling technology for “Infrastructure as Service” provisioning, increasing the flexibility and cost savings offered to customers. By extending concept of server network infrastructure, allocation different, independent virtual networks over single physical carried out on demand. A fundamental challenge in systems choose which nodes links use hosting known “virtual allocation” problem. All proposals elastic optical assume centralized operation, deploying node...