A Natural Language Parameter Catalogue for Algorithm-Aided Design of Modular Housing

Abstract

The construction industry is embracing digital technologies, particularly generative or algorithm-aided design principles. However, integrating these digital tools into design processes while ensuring compliance with rules and regulations remains a significant challenge. This study aims to address this challenge by identifying the essential design parameters and constraints required for semi-automated building model design in the early design stages, with a specific focus on a use case of modular multi-story housing in Vienna. To achieve this, we investigate which parameters are fundamental, how constraints must be formulated, and how these aspects can be covered effectively in the design process. Our research provides a parameter catalog in natural language format to be used for scripting algorithms to generate parametric models. We delineate crucial housing-specific design parameters and identify constraints derived from legal, technical, evaluative, and expert knowledge sources. These constraints ensure that the designs comply with regulations and standards. The findings are organized into eight thematic clusters, each detailing specific conditions, and their interdependencies, thus offering a guideline for scripting algorithm-aided design processes. In conclusion, we propose a conceptual model for translating natural language design parameters into visual programming language.