Intelligent Energy-Optimized Resource Allocation With QoT Consideration in Hybrid C/C+L Optical Networks

Abstract

With the explosive growth of network traffic driven by emerging applications such as 6G and AI, traditional C-band optical networks are approaching their 4.8 THz capacity limit. Although extending to the C+L band provides up to 12 THz of spectrum resources, full-band deployment inevitably leads to increased costs and energy consumption. In the context of the urgent demand for green and low-carbon development, optimizing energy consumption while ensuring transmission quality has become a critical challenge in the evolution of optical networks. To address this issue, we construct an energy consumption model for hybrid C/C+L-band optical networks, quantitatively characterizing the static and dynamic power consumption of transceivers, optical amplifiers, and optical cross-connects. On this basis, we establish an energy minimization mathematical framework that evaluates QoT through GSNR calculation, ensuring accurate performance estimation under different modulation formats. Furthermore, we propose a joint optimization algorithm based on the Deep Deterministic Policy Gradient (DDPG-EORA), which covers routing, band allocation, modulation format selection, and spectrum assignment. Extensive numerical results demonstrate that DDPG-EORA can effectively reduce both network energy consumption and blocking probability, while achieving superior coordination between energy efficiency and QoT assurance, thereby supporting request transmission even under high traffic loads.