INTRODUCING MACHINE LEARNING WITHIN AN INTERACTIVE EVOLUTIONARY DESIGN ENVIRONMENT
The present work focuses on providing machine based support to the designer within an interactive evolutionary design environment (IEDE). This improves the interactivity of the IEDE by reducing the cognitive load placed on the designer due to repetitive assessment of solutions. The background for this work is an interactive evolutionary bridge design system which takes into account rule based and subjective (i.e. designer based) aesthetic fitness along with engineering fitness of the solutions. The machine learning system attempts to learn the user’s aesthetic preferences and thus takes over the subjective evaluation of solutions. Fuzzy rule based, case based and radial basis function based techniques are applied to the problem.