Low-complexity optical fiber optimization design method

Abstract

The invention relates to a low-complexity optical fiber optimization design method, which comprises the steps of S1, determining a global optimal neighborhood positioning algorithm, and positioning aglobal optimal neighborhood based on an adaptive search decision of a function gravity center of the algorithm; S2, constructing and training a BP neural network model; and S3, in the global optimal neighborhood range obtained by the first optimization, calculating based on the trained BP neural network to obtain a corresponding data set, calling the global optimal neighborhood positioning algorithm again for the data set to perform secondary optimization, and searching and outputting optimal optical fiber structure parameters under the conditions of actual optical fiber drawing and application limitation. According to the method, a global optimal neighborhood positioning algorithm is utilized to reduce the modeling range of the BP neural network, so that the modeling precision is improved, and the precision of optimal performance prediction in the optical fiber optimization design process is improved.

Description

technical field [0001] The invention relates to the technical field of optical communication, and more particularly, to a low-complexity optical fiber optimal design method. Background technique [0002] In recent years, the further expansion of the existing commercial single-mode optical fiber system has begun to be limited, and the capacity of its optical fiber communication system is gradually approaching its theoretical limit. Space-division multiplexing optical fiber communication system has received extensive attention in recent years, mainly two types of few-mode optical fiber and multi-core optical fiber. For the design of new and different optical fibers, the existing optical fiber designs are mostly based on repeated electromagnetic calculations or testing methods. It is often time-consuming and complex to find the optimal performance optical fiber design under specific conditions. [0003] In addition, there are methods that use machine learning to assist optical...

AI Technical Summary

Problems solved by technology

However, the modeling of this method is to achieve accurate neural network modeling with more input and less output in a small range in a priori specific circumstances, and in more general cases, first, there is no reliable prior Under certain conditions, the design parameters of the optical fiber should be complex nonlinear relationships in a large range; second, the modeling problem of less input and more output that will be encountered in the design of some optical fibers is not considered; third, the accuracy of the optimized design results will be reduced It completely depends on the prediction accuracy of the neural network; combining three general problems, in the optical fiber optimization method based on direct neural network modeling, it will appear that the neural network design is difficult, the training complexity and time complexity are high, and the relative prediction error is relatively small. Large, so it is difficult to guarantee the prediction accuracy, and it does not have universal applicability

[0004] In addition, it is also proposed to use intelligent optimization methods such as genetic algorithm (GA) to perform multi-point detection on the optimization space in parallel to carry out optical fiber optimal design, but the calculation amount is large and the convergence speed is slow, which often occurs for more complex models. The premature convergence phenomenon does not guarantee finding the optimal

Images