Structural Topology Optimization of Magnetic Actuators Using Genetic Algorithms and ON/OFF Sensitivity

In genetic algorithm (GA) based topology optimization problems, characteristics of an initial population are important for the rapid and stable convergence. This paper introduces an algorithm generating randomly an initial population with superior hereditary characteristics. To avoid the generation of small structural spots, the blurring technique is proposed. The connectivity of seed elements considerging the magnetic flux flow in the design domain is focused. Differently from the classical GA by linear strings, this study deals with two-dimensonal chromosomes and the geographic crossover method to increase the diversity of offspring. The proposed design algorithm is applied to the yoke optimization of magnetic actuators.