A broad and narrow approach to interactive evolutionary design—An aircraft design example

While interactive evolutionary computation (IEC) is starting to penetrate a larger scientific community, only few researchers have applied IEC to the design of complicated artifacts like machines or transportation systems. The present paper introduces a specific approach to interactive evolutionary computation that breaches the two historical categories of user-defined fitness and selection in each generation (narrow) and occasional user-intervention of an automated evolutionary process to correct the fitness function used for (multi-objective) optimization (broad). To highlight the approach, a real world aircraft design problem is employed that demonstrates the relevance and importance of both features for an effective design process.