Machine Learning Approach for Computing Optical Properties of a Photonic Crystal Fiber

Photonic Crystal Fibers (PCFs) find critical applications in various fields, such as telecommunications, optical sensing, and nonlinear optics, because of their remarkable optical properties. However, typical numerical simulations, like the one conducted with COMSOL Multiphysics, are computationally intensive and time-consuming, preventing them from achieving scalability and real-time utility. In response to these shortcomings, this project proposes a machine learning (ML)-based framework for the efficient prediction of several important PCF optical characteristics, with an initial focus on calculating confinement loss and sensitivity for sensor applications. Using-datasets generated from simulations, using data preprocessing methods, and implementing numerous ML algorithms, such as regression models, decision trees, bagging, and gradient boosting, the developed system or model generates highly accurate,almost instant predictions without needing repeated complex simulations.This new method drastically cuts down computational expenses, reduces the use of high-end hardware, and allows PCF scalable design exploration. By integrating machine learning with traditional simulation techniques, this work enhances the efficiency of photonic research, accelerating advancements in optical design and more other scientific and research application.By combining the gap between artificial intelligence and computational photonics, the new technique enables faster results and optimal innovation in PCF structures. Its flexibility also makes it a powerful tool for advancing integrated photonics and next-generation optical technologies.