Preprocessing convolutional neural networks based routing algorithm for PILOSS optical switching networks

For path-independent insertion-loss (PILOSS) optical switching networks, the traditional XY routing algorithm fails to ensure high-quality end-to-end communication, highlighting the need for a more efficient routing algorithm. This paper introduces a preprocessing convolutional neural networks (CNN) based routing algorithm for PILOSS optical switching networks. The proposed routing algorithm addresses the issue of partial permutation when CNN is directly applied to the routing tables. The routing algorithm is redefined as a classification task, with the CNN monitoring multipath optical power and providing three classification references. The results show that CNN’s prediction accuracy improved from 47.56% to 90% after preprocessing. In optimal conditions, using 30G PAM4 modulation and with a fixed symbol error rate (SER) of 10 −9 , the lowest received optical power is -8.31dBm, nearly equivalent to the back-to-back (B2B) received power. Furthermore, crosstalk ranged between -50.31dB and -22.70dB, significantly outperforming the traditional XY algorithm.