ABSTRACT
This chapter illustrates a novel framework combining building information model (BIM) and various artificial intelligence techniques developed as part of The Pennsylvania State University’s participation in NASA’s 3D-Printed Mars Habitat Centennial Challenge. It encompasses a parametric algorithm for generating design alternatives and a common BIM-based data exchange platform facilitating information exchange between the parametric algorithm, analysis tools, 4D-simulation tools, and optimization routines. Machine learning is embedded to learn from previous design iterations and speed up the optimization process. This framework enables generating 3D models of multiple design alternatives, 4D simulations of robotic additive construction, and design optimization for structural and environmental performance. Simulating construction early in the design phase is important as it plays a crucial role in dictating the constructability of the building shape and, therefore, the feasibility of autonomous construction. The framework also overcomes limitations of designers in generating and evaluating large sets of building design alternatives and it is demonstrated in the design of a 3D-printable habitat. Limitations of the proposed framework and potential avenues for addressing those limitations are outlined.
