Skip to the content.

Generative Models for New Town Location and Design

Project ID: 2531ad1497

(You will need this ID for your application)

Research Theme: Artificial Intelligence and Robotics

UCL Lead department: Centre for Advanced Spatial Analysis (CASA)

Department Website

Lead Supervisor: Batty Michael

Project Summary:

New towns lie at the heart of the physical planning process and along with greenbelts represent powerful instruments for developing new communities on greenfield sites and brownfield land. These have the potential to address the current housing crisis, the regeneration of old infrastructures particularly railways, and the agenda for growth.

In the last 100 years, nearly 50 new towns have been developed in the UK and we are about to begin a wave of new town building for which we require the best locations that will optimise environmental quality, accessibility, physical topographies etc. thus avoiding uncoordinated urban sprawl.

This project will develop methods for identifying these best locations, building on overlay analysis, geodesign, and various weighting procedures, thence generating data and their relationships that enable us to rapidly screen locations across the entire UK. These methods integrate design networks with machine learning, and once developed can be generalised to many different locational problems.

The project will produce methods and software which can be used by the New Towns Commission to which we are linked through the Digital Task Force for Planning. The software will be in the public domain and will build on methods already begun as illustrated in the paper: Batty M, et al 2024 Participatory design based on opinion pooling. Philosophical Transactions of the Royal Society A 382: 20240108.