Urban Decision-Making and Complex Systems

UDM_Singapore 2011

Diorama of the city of Singapore used for planning purposes. Singapore, September 2011.

How can technology assist in decision-making regarding such things as urban planning and risk analysis?

This is a cross-disciplinary program to use and develop advanced technologies in areas such as agent-based simulation to support urban decision-making.

Research focus

Many decisions in planning for our cities, involve an understanding of complex interactions between different aspects of the city—from its infrastructure, its buildings to its inhabitants and culture. This program focuses on leading-edge information technology and techniques, and how they can be applied to specific questions and issues in urban decision-making. One specific focus of the program is simulation, in particular agent-based simulation. There is a particular focus on developing a platform that supports integration of separately developed simulation modules within a larger whole, as well as re-usability of modules where possible.

Description of program

The program seeks to develop general approaches and tools that can be applied to a range of urban decision-making issues and questions. This involves the development and exploration of technologies and techniques to assist urban decision-makers in understanding the complex systems in which they are developing policies and infrastructure. The program aims to take specific questions and issues, and explore the use of leading-edge technology to contribute to addressing these questions and issues, with a view to building strong expertise in the interdisciplinary space between the social sciences and technology.

Research themes

The overarching focus of the program is to look at how advanced technology can support decision-making and risk analysis in complex urban systems. The current themes are:

1. Agent based simulation to support risk assessment, policy and planning
Agent-based simulation is a powerful tool for developing understanding of complex, multi-scalar, multi-actor systems. We are exploring a range of issues that will make this technology both more accessible to end-users, and also more reliable and more transparent. In order for government departments and other groups to be confident in using this technology, it is important that they are able to understand the underlying models and re-use models or systems that they already have and trust. It is also important to develop scientific understanding of how to use simulations to systematically explore risks and vulnerabilities, to gain a nuanced understanding of the system under consideration. It is impossible to explore all possible scenarios—but it is important to explore both a sufficient number, and sufficient combinations of key aspects. We are developing ways to support end-users in obtaining a principled set of simulations on which they can reliably base their understanding.

2. Participatory modelling of complex systems
Any computational system is only as good as the underlying model which it is based on. In modelling complex systems it is crucial to identify the key components of the system and how they interact. There is broad support for the idea that this is best done by actively involving the end-users and the subjects being modelled. One of the research themes of the program is how to best support participatory modelling, using the skills of a multi-disciplinary team involving both computer scientists and social scientists, as well as end-users of systems.

3. Computational models of human behaviour
One of the difficulties in agent-based modelling is that the approach to modelling humans and their behaviours is very simplistic, and is essentially just one off reactive rules. Social scientists typically find this far too restrictive for believable modelling of humans, leading to lack of confidence in simulation results. We are exploring integration of more sophisticated behaviour modelling, incorporating pursuit of multiple goals over time. We are also looking at how social science theories can be made sufficiently simple and precise (while still capturing essential elements) for implementation in a computational system. We will then explore to what extent this assists in facilitating the use of agent-based simulation technology in policy and planning situations.


For more information on Urban Decision-Making and Complex Systems, please contact the Program Leader, Lin Padgham.