Imagine living in a peaceful, green city with parks and sidewalks, bike lanes and buses that take people to shops, schools and service centers within minutes.
That airy dream is the epitome of urban planning, embodied in the idea of a 15-minute city, where all basic needs and services are accessible within a quarter of an hour, improving public health and reducing vehicle emissions.
Artificial intelligence could help urban planners realize that vision faster, a new study from researchers at Tsinghua University in China shows how machine learning can generate more efficient spatial layouts than humans, and That too in less time.
Automation scientist Yu Zheng and colleagues wanted to find new solutions to improve our cities, which are becoming increasingly congested and concrete.
They developed an AI system to tackle the toughest, computational tasks of urban planning – and found that it produced urban plans that outperformed human designs by about 50 percent on three metrics: access to services and green spaces, and traffic levels.
Starting small, Zheng and colleagues tasked their model with designing urban areas of only a few square kilometers (about 3×3 blocks) in size.
After two days of training and using multiple neural networks, the AI system discovered the ideal road layout and land use to fit with the 15-minute concept of city and local planning policies and needs.
While Zheng and colleagues’ AI model has some features to enhance its use for planning large urban areas, designing an entire city would be infinitely more complex. Researchers estimate that designing a neighborhood consisting of 4×4 blocks involves twice as many planning decisions as 3×3 blocks.
But automating even a few steps in the planning process can save huge amounts of time: AI models calculate in seconds some tasks that would take human planners 50 to 100 minutes to complete.
Researchers say that automating the most time-consuming tasks of urban planning would free up planners to focus on more challenging or human-centered tasks, such as public engagement and aesthetics.
Rather than AI replacing people, Zheng and colleagues envision their AI system serving as an ‘assistant’ to urban planners, who can generate concept designs optimized by algorithms, and based on community feedback. But can be reviewed, adjusted and evaluated by human experts.
This final step is central to good design, writes Paolo Santi, a research scientist at the Massachusetts Institute of Technology (MIT), in a commentary on the study.
Urban planning “is not simply the allocation of space for buildings, parks and functions, but the design of a place where urban communities will live, work, interact and, hopefully, thrive for a very long time,” he writes.
Comparing their human-AI workflow to human-only designs, Zheng and colleagues found that the collaborative process could increase access to basic services and parks by 12 and 5 percent, respectively.
The researchers also surveyed 100 urban designers who were unaware whether the plans they were asked to choose between were drawn by human planners or AI. AI won significantly more votes for some of its spatial designs, but for other plans, there was no clear preference among survey participants.
Certainly the true test will be in the communities built from those plans, measured by the reductions in noise, heat and pollution, and improvements in public health, that better urban planning promises to bring.
The study has been published in Nature Computational Science.