Efficient segmented power control for C+L+S ultra-wideband fiber transmission using CMA-ES

Area of Science:

• Optical Communications
• Signal Processing
• Fiber Optics

Background:

• Ultra-wideband (UWB) fiber transmission systems (C+L+S bands) require precise launch power control.
• Inter-channel stimulated Raman scattering (SRS) and bandwidth utilization are key challenges in UWB systems.

Purpose of the Study:

• To investigate efficient launch power optimization strategies for UWB fiber systems under per-band constraints.
• To compare the performance of Covariance Matrix Adaptation Evolution Strategy (CMA-ES) with Particle Swarm Optimization (PSO).
• To evaluate optimization objectives including Signal-to-Noise Ratio (SNR) flatness and total throughput.

Main Methods:

• Application of a segmented slope-offset model for power optimization.
• Utilization of CMA-ES for optimizing launch power distribution.
• Numerical simulations to assess SNR flatness and channel capacity.

Main Results:

• CMA-ES demonstrated comparable optimization performance to PSO with significantly faster convergence.
• Segmented power optimization improved SNR flatness to 0.335 dB ripple while maintaining per-channel rates above 800 Gb/s.
• Capacity maximization reached 161.9 Tb/s, with diminishing returns from further segmentation.

Conclusions:

• CMA-ES is an effective and efficient method for power optimization in UWB fiber systems.
• Segmented power optimization is beneficial for SNR flatness, while capacity maximization shows limited gains from extensive segmentation.
• The triangular approximation for power distribution is only suitable for capacity maximization objectives.