Application of Preference Information in Truss Design

Many automated structural design methods have been developed to improve design efficiency and quality. However, most of them can only consider the quantitative objectives and constraints. This may ignore the important qualitative design information, which limits the functionality and efficiency of these methods. To solve this problem, qualitative information needs to be effectively integrated into the design process. To this end, this study aims at modeling the user shape preference to guide the generation of satisfactory truss structures. A prediction model for user preference is proposed by generating the data set and selecting the appropriate machine learning method. Besides, Physical Programming is adopted to combine the prediction model with a grammar-based computational design method. Finally, a new automated design method for truss structures is developed.