Optimization Of Photonic Crystal Fiber Dispersion Properties Using Genetic Algorithm

The researches of Photonic Crystal Fiber (PCF) have been blooming for years in thefields of optoelectronic devices. Besides the optical mechanism and the manufacturingmethods, dispersion is also a significant and sophisticated parameter. It is not only themain reason for pulse broadening which results in signal distortion and limits the trans-mission speed in optic communication, but also an essential part which has direct impactoncross-phasemodulationandfour-wavemixinginnonlinearoptics. Akindofoptimiza-tion algorithm named Genetic Algorithm is introduced while the effective index methodand the finite element method are also used for simulation, and the reverse optimizationof the dispersion properties is fulfilled. The major content of this thesis includes:First, the fundamental theory and the application fields of EIM and FEM are intro-duced. The analysis of the four feature properties of GA: encode, group search, fitnessevaluation and parallelism shows that it is suitable and effective to solve the nonlinear re-verseproblemofPCFdispersionoptimization. Theflowchartoftheoptimizationprocessis illustrated when GA and EIM and FEM are combined respectively. The advantages ofmixing programming of FORTRAN and MATLAB are discussed and the GUI version ofthis program is accomplished.Second, based on the theory and program mentioned above and with self-adaptivefundamental mode search method, PCFs structure design with zero dispersion from visi-ble to mid-infrared, and flat dispersion cover S, L and C communication band and largenegative dispersion are fulfilled.Finally, a terahertz PCF design with low loss, high birefringence0.012at0.3THzis completed. The research results are as follows: the increase of the space between theair holes and the air filling factor would shift the zero dispersion wavelength to lowerfrequencies; high birefringence could be introduced by alter the air holes sizes in twodirections which breaks the six fold symmetry; there would be little difference betweenthe two dispersion in different polarized direction as the confinement is different at thesame frequency.In conclusion, the nonlinear reverse optimization of PCF dispersion properties couldbe solved efficiently and precisely by GA combined with EIM and FEM. The theory andresults would provide valuable reference for PCF design.