摘要

In the architectural design process, the conceptual design stage is to devise a creative alternative in response to the intent of the architects. In this paper, we propose an alternative evolutionary-based architectural design method by using the Implicit Redundant Representation Genetic Algorithm (IRRGA) that is highly suited to explore unstructured problem formulation such as conceptual design. Also, a new string representation for apartment building designs is proposed such that the size and the number of apartment units with stairs are not fixed and can be changed during the design evolution. The design objectives are selectively applied in terms of symmetry, structure, circulation, and facade. Each objective is used respectively as a fitness function to demonstrate the performance of IRRGA. Finally, a multi-objective fitness function is applied and the resulting apartment building designs show their own level of creativity.

  • 出版日期2016-12