Electromagnetic Device Optimization using Improved Differential Evolution Methods

Abstract

Recently, evolutionary algorithms (e.g. genetic algorithms, evolutionary programming, and evolution strategies) have proven to be useful tools for the optimization of difficult problems in electromagnetics. Differential evolution (DE) is one comparatively simple variant of an evolutionary algorithm using floating-point encoding and few control parameters. This work presents improved DE algorithms based on linearly time varying control parameters, sinusoidal functions, and diversity analysis of population