Structural form generation using interactive genetic algorithm

Structural form design could be considered as a bi-objective problem which should satisfy both the efficiency criterion of structural engineering and the aesthetic criteria of architects. This research tries to introduce Interactive Genetic Algorithm (IGA) in the problem of structural form design. It combines the structural analysis performed by computer and aesthetic evaluation by architects into a bi-objective IGA process, in order to generate structural forms which are preferred by the architects and at the same time structurally optimal. In this research, the structure generated consists of two kinds of members, truss and beam. Generation and evolution of structure is based on a triangular element composed of several members. Through experiment of the IGA structural form design system, it is found the structure forms are optimized as the evolutionary process proceeds, and the aesthetic preference of architect is also transferred from generation to generation. It is also revealed that the two criteria have mutual restrictions, which resulted in compromised results.