A Data-driven Optimization of First-order Regular Perturbation Coefficients for Fiber Nonlinearities

TL;DR

This paper uses gradient descent to optimize first-order regular perturbation coefficients in fiber nonlinearities, improving model accuracy and simplicity.

Contribution

It introduces a data-driven optimization method for FRP coefficients, extending validity and reducing complexity compared to traditional approaches.

Findings

Optimized coefficients extend the FRP model's validity range.

The optimized model reduces complexity while maintaining accuracy.

Gradient descent effectively estimates coefficients for fiber nonlinearities.

Abstract

We study the performance of gradient-descent optimization to estimate the coefficients of the discrete-time first-order regular perturbation (FRP). With respect to numerically computed coefficients, the optimized coefficients yield a model that (i) extends the FRP range of validity, and (ii) reduces the model's complexity.