Research On Arbitrary Mode Conversion Based On Intelligent Algorithm In Space Division Multiplexing Optical Fiber Communications

In recent years, with the explosive growth of network traffic, time division multiplexing(TDM), wavelength division multiplexing (WDM),polarization multiplexing and high-order modulation technologies, once solved the network traffic demands at the advantages of fully using the degrees of freedom including time, wavelength, polarization, amplitude and phase. However, this single-mode fiber's nonlinear limits the further development of network traffic. Space as a new dimension, its derived space-division multiplexing (SDM) technology is one of the most promising approaches to break this limitation. In this paper, a series of intelligent algorithms are proposed to realize the conversions of high precision mode.The innovation and main work are as follows:1. A mode conversion based on multi-phase intelligent algorithm is proposed.In this paper, multi-phase is applied to the mode conversion, and multi-phase simulated annealing algorithm (SA) and multi-phase genetic algorithm (GA) are proposed. The paper also introduces their realization principles and simulation results. Compared with binary-phase intelligent algorithm, the accuracy of conversion we proposed increases from 80% to 99%or more, achieving a true sense of high-precision mode conversion.2. Local adaptation is proposed to optimize the efficiency for mode conversionLocal adaption refers to the mutual conversion between different modes,which dynamically select of annealing range based on the order of converted modes. More specifically, different conversion modes have different state of search space. This method based on intelligent algorithm can be applied to the mode conversion, which eliminates the influence on search efficiency resulted from the useless state space and greatly improves the efficiency of mode conversion process.3. A genetic heuristic search algorithm is proposedSimple traditional genetic algorithm has two fatal shortcomings: low search efficiency on both early and late genetic state, and slow convergence rate resulted from more than one generation of population stagnation. In this paper, the multi-phase simulated annealing algorithm is applied into the genetic operators, proposing the annealing, cross-annealing and mutation annealing operators. The low search efficiency of the late genetic state is improved, and the advantages and disadvantages of the three algorithms are analyzed.4. The experiment verifies the feasibility of the intelligent algorithm in mode conversionIn this paper, an experimental platform of mode conversion is constructed. The model transformation based on multi-phase simulated annealing algorithm is applied to the experiment, and generated high accuracy results, which verifies the effectiveness of the algorithm.