Comparison of speed and accuracy of optimization algorithms for fiber optic communication lines

In the pursuit of optimizing optical signal transmission quality, a critical concern is the surplus of the signal-to-noise ratio over its threshold, known as the margin. This metric reflects communication reliability and transmission efficiency. Optimization comes down to finding the extremum of a function of many variables, unique for each fiber-optic communication line. This paper systematically examines various optimization algorithms to identify the most suitable for this diverse function set. Bayesian optimization-based algorithms are particularly scrutinized. Empirical analysis demonstrates reduced computational time compared to deterministic global optimization methods, while maintaining an acceptable deviation from the optimum. This research contributes to improving efficiency and reliability in fiber-optic communication systems, enhancing signal transmission quality optimization.