ON THIS PAGE

AI-Assisted Green Space Layout Optimization for Smart-City Environmental Design

Patsy Healey1, Kongjian Yu2, Xianheng Zheng2
1Patsy Healey, Newcastle University, King’s Gate, Newcastle upon Tyne, NE17RU, UK
2College of Architecture and Landscape Architecture, Peking University, Beijing 100871, China

Abstract

As ecological principles increasingly guide contemporary architectural and urban design, optimizing green-space layouts has become central to improving built-environment performance. This study applies the Pix2Pix model to interpret planning-area maps and develops a green-space layout support system implemented in a Unity3D engine with a Python-based workflow. The system enables visual simulation of alternative green-layout schemes and helps designers compare options to identify a preferred configuration. In a case validation, the AI-assisted optimization adopts a zoned daylighting strategy for public areas. Simulation results show reductions in district-building energy use, with cooling demand decreasing from 75.64 kWh/m² to 65.32 kWh/m² and heating demand dropping from 45.26 kWh/m² to 42.31 kWh/m². Residents also reported high satisfaction with the simulated retrofit outcomes. Overall, the proposed approach supports environmentally responsible and sustainable smart-city development by lowering environmental impacts and contributing to healthier, more livable urban spaces.

Related Articles
Andreas Faludi1, Peter Nijkamp1, Jaap Evert Abrahamse1
1Department of Urbanism, Faculty of Architecture, Delft University of Technology, Delft, Netherlands
Anne Vernez Moudon1, Nabeel Irfan1, Steven Robbins1
1Department of Urban Design and planning , University of Washington, Seattle, WA 98105, United States
Yan Li1, Xiaong Wie1
1Anhui Vocational and Technical College, Hefei, Anhui, 230011, China
Jie Tian1, Dongdong Gao1
1Deaprtment of Urban Design and planning , University of Washington, Seattle, WA 98105, United States