Development of photonic devices using Genetic Algorithm integrated with Shannon entropy

Abstract

Genetic Algorithm has become a powerful metaheuristic method for developing photonic devices. Metaheuristic optimization allows challenging designs when associated with numerical methods of electromagnetic analysis. However, populations can become large, which increases total optimization time. Based on Shannon entropy, this article presents a novel methodology that acts on the distribution of initial solutions in the search space. Hence, the proposed method can benefit the optimization processes of photonic devices, consequently accelerating the convergence. The study shows three applications of this technique, both of which exhibited improved convergence optimization.