From Linear to Recursive: An AI-Based Educational Model for Architectural Programming

From Linear to Recursive: An AI-Based Educational Model for Architectural Programming

Kamran Javaheri11 Hana Avani2

1) Department of architecture, Se.C, Islamic Azad University, Semnan, Iran,
2) Department of architecture, Se.C, Islamic Azad University, Semnan, Iran,

IOI Code : AAHD-ABGAHA
Publication : National Conference on Information Technology, Nanotechnology, Artificial Intelligence and Technological Futures Studies(itnaf.ir)
Abstract :
    Architectural programming, as the first and most determining stage of the design process, has historically been dominated by linear, step-by-step models. This linear approach, when confronted with the complexities of contemporary architectural projects—simultaneously engaging aesthetic, functional, environmental, and social variables—fails to establish dynamic feedback loops and learning from subsequent design decisions. This study aims to design and validate an AI-based educational model that transforms architectural programming from a linear process into a recursive one with intelligent feedback loops. A mixed-methods approach with a sequential exploratory design was employed. In the qualitative phase, content analysis of specialized texts and semi-structured interviews with experts in architectural education and physical programming were conducted. In the quantitative phase, the proposed model was developed as an agent-based intelligent system and subsequently implemented in an actual educational workshop with architecture students. The findings revealed that the intelligent recursive model, through generating constructive contradictions, predicting secondary consequences, and learning from the path rather than the destination, significantly improves the quality of architectural programs and students learning experiences. The five key components of the proposed model are: inherent instability of assumptions, constructive contradictions, prediction of secondary consequences, learning from the path rather than the destination, and recursive self-efficacy. The three-layered model derived from this research (physical → behavioral → semantic) provides a theoretical framework for understanding and designing vitality in residential collective spaces, demonstrating that successful design is one that simultaneously attends to what is built, how it is used, and how it is experienced. By introducing the concept of recursive flexibility as a new indicator for assessing the quality of architectural programs, this research expands the scope of evaluation criteria.

Keywords : architectural programming recursive model artificial intelligence intelligent feedback loop design education collective spaces vitality