Strategies for building spatial and structural design generation and optimisation

Two strategies are presented that generate an optimised structural design by topology optimisation, which is applied in the first strategy to specific areas in a given spatial design and in the second strategy to the complete spatial design. Two other strategies apply evolutionary principles to generate optimised sets of one spatial design and one accompanying structural design via either a co-evolutionary method or a genetic algorithm. All four strategies are compared using academic spatial designs, and the co-evolutionary strategy is compared with a case study as well. The efficacy of topology optimisation depends on the load type; topology optimisation generates complete 3D building structural designs; and modifying the spatial design may be as effective as topology optimisation. Compared to the co-evolutionary method, the genetic algorithm finds more design alternatives, but converges slowly, however outcomes of both methods are similar. The case study shows that the co-evolutionary strategy provides useful design alternatives, yet is not capable of addressing the level of detail found in practice. Guidelines for the practical and academic applications of the strategies are given