Performance-based Generative Architecture Design: A Review on Design Problem Formulation and Software Utilization

Performance-based Generative Architecture Design (PGAD) is a design approach based on coupling parametric modelling with performance simulation and optimization tools. It is highly efficient and advanced but design problems, because of their many complex issues, are often ill-defined and must be somehow preprocessed and streamlined. References are collected from paper indexing databases and are analyse with the software, Citespace. When it comes to detailed contents, comparison and clustering are conducted to illustrate how design problems are formulated among research papers. In the literature, five categories of building components are frequently pursued. To obtain these building components by the PGAD method, there are eight types of relationships among building components, decision variables, performance objectives, and performance indicators that should be contemplated. After exposing these correlations, various implementation software and tools for realizing the PGAD are compared. Four types of integrated design environment are identified and compared based on their advantages and disadvantages. This study aims at investigating the status quo of the topic and promoting PGAD development.