An ML model based on precomputed per-channel SCI is proposed. Due to its superior accuracy over closed-form GN, an average SNR gain of 1.1 dB in an end-to-end link optimization and a 40% reduction in required lightpaths to meet traffic requests in a network planning scenario are shown.
翻译:提议采用以预先计算过的每个SCI频道为基础的ML模型,由于该模型的精度高于封闭式GN,显示SNR在端到端链路优化中平均增加1.1 dB,在网络规划设想中,为满足交通需求所需的光路减少40%。