Evolutionary Design of Optical Waveguide with Multiple Objectives

In this paper, we investigate a real-world problem, constructing optical waveguide structures using evolutionary search strategies. Optical waveguide is the most basic component in optical communication and integrated optical circuits. The structure of a waveguide is of great importance as it would significantly impact on the quality of light transmission. The aim of this paper is to find a set of potential structures which satisfy multiple waveguide design objectives including minimum group velocity dispersion and minimum propagation loss. Therefore, evolutionary algorithms which are populate-based search techniques are more suitable for this type of tasks. As a part of this investigation, a GP-based parametric optimization methodology called Parameter Mapping Approach (PMA) is introduced. This method together with traditional GA have been adapted into this study. The experiment results demonstrate both PMA and GA can produce multiple waveguide structures that meet the design criteria. Furthermore these evolved structures have very low dispersion and loss compared to those reported in the current literature.