Stochastic searching algorithms such as the Genetic Algorithms (GA's) are commonly used for shape optimization of electromagnetic devices. Although very robust and easy applicable, the GA's are usually time consuming. In this paper, we present a new improved method for optimization of electromagnetic devices that utilizes the ordinary GA with enriched searching population of so called assistant chromosomes. These new assistant chromosomes are generated in accordance with the values of the objective function. The proposed procedure is highly flexible and exhibits improvements of the searching speed and accuracy of the computed results.
展开▼