Comprehensive literature review: advancing sustainable urban environments through machine learning usage in digital architecture

This research delves into the revolutionary role of machine learning (ML) in digital architecture, with its use in the development of sustainable urban spaces. With urban cities grappling with increased concerns over energy efficiency, climate resilience, and resource optimization, ML has proven to be an instrumental aid in the design of data-led, responsive architectural solutions. A systematic review was performed on 43 peer-reviewed papers to analyze ML methods such as deep learning, generative design systems, and metaheuristic algorithms, for optimizing urban planning, energy management, and building performance. The results suggest an increasing use of ML toward sustainable design and highlight interdisciplinary work and standardized frameworks as success indicators. The research framework identifies three fundamental phases: (1) organizational and technological inputs, (2) mediating factors focusing on data-driven design and interdisciplinary collaboration, and (3) the anticipated outcomes of integrating ML into sustainable architecture. Despite this, factors like data quality, regulatory hurdles, and policy support are obstacles to its mass implementation. This paper builds on current research by offering a systematic approach to embedding ML within architecture, calling for experimentation in real-world contexts and interdisciplinary cooperation. Research in the future must work to develop ML models with better datasets, regulation consistency, and architectural training using AI.