Abstract
Recently, the design and fabrication of optical devices become more and more complex. Through numerical simulation and analysis for linear and nonlinear optical devices, we can reduce the fabrication cost dramatically. In this work we use genetic algorithm (GA) realize the optimization for linear and nonlinear optical devices. With the ability for finding a global optimal solution by the GA method, we hope to solve the problems of inverse design in optical devices optimization. For the linear optical devices, we use Layer-Peeling algorithm both for designing single and multi-channels optical fiber Bragg gratings with rectangle reflection filters. Especially for multi-channel fiber Bragg gratings, first we use a Sinc function as the sampling functions in the index modulation profiles. Then GA method is used to optimize the relative phases in each channels. And for the nonlinear optical devices, we use GA method to synthesize quantum-phase-matching (QPM) devices with the considerations of constrained conditions in practical fabrications. A QPM device with two Gaussian profiles in the transmission spectrum is demonstrated.