While the hanging model and physical prototyping are reliable methods for form-finding continuous lightweight shell structures, they cannot explore various design possibilities and consider multiple criteria in the design. This research aims to employ metaheuristic algorithms as a search tool for simultaneous form-finding (form exploration) and optimization, known as computational morphogenesis. This workflow consists of a parametric model and metaheuristic algorithms, which were implemented through costume Python codes. To evaluate the workflow, a double-curvature footbridge was defined, and the forms that GA found were in line with the anticipation. The GA found a better solution (0.04022) than the SA (0.064) by considering elastic energy change. In case study II, the optimal form of a continuous lightweight shell was explored by utilizing GA, PSO, SA, and the hybrid algorithm of GAPSO. The GA improved the fitness solution by 35% in the second run, while PSO showed similar results, and the hybrid GAPSO algorithm did not yield better results. The results show that the optimum result based on different algorithms was not significantly different. Nevertheless, hyperparameters considerably affected finding the global minimum. Moreover, NSGA-II was utilized, which let the designer trade-off between alternatives while considering other criteria.
Read full abstract