Skip to content

Data Assimilation for Agent-Based Models - A research project at the University of Leeds, funded by the European Research Council

License

Notifications You must be signed in to change notification settings

Urban-Analytics/dust

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

dust

Main dust website: https://urban-analytics.github.io/dust/

The DUST project is an initiative at the University of Leeds that has been funded with €1.5M from the European Research Council. It started in Janurary 2018 and was featured as one of the Research Council’s highlighted projects in a recent press release. For more information please contact the Principal Investigator: Nick Malleson (http://www.geog.leeds.ac.uk/people/n.malleson/).

Project Results

The project ended in December 2023. If you would like to learn more about what the project did, the best places to look are on the publications and presentations pages of the main project website.

Most of the project code is in this repository, but there are some other related repositories as well, particularly looking at the footfall analysis work that we did in the final stages:

Project Aims

Civil emergencies such as flooding, terrorist attacks, fire, etc., can have devastating impacts on people, infrastructure, and economies. Knowing how to best respond to an emergency can be extremely difficult because building a clear picture of the emerging situation is challenging with the limited data and modelling capabilities that are available. Agent-based modelling (ABM) is a field that excels in its ability to simulate human systems and has therefore become a popular tool for simulating disasters and for modelling strategies that are aimed at mitigating developing problems. However, the field suffers from a serious drawback: models are not able to incorporate up-to-date data (e.g. social media, mobile telephone use, public transport records, etc.). Instead they are initialised with historical data and therefore their forecasts diverge rapidly from reality.

To address this major shortcoming, this new research project will develop dynamic data assimilation methods for use in ABMs. These techniques have already revolutionised weather forecasts and could offer the same advantages for ABMs of social systems. There are serious methodological barriers that must be overcome, but this research has the potential to produce a step change in the ability of models to create accurate short-term forecasts of social systems.

The project will evidence the efficacy of the new methods by developing a cutting-edge simulation of a city – entitled the Dynamic Urban Simulation Technique (DUST) – that can be dynamically optimised with streaming ‘big’ data. The model will ultimately be used in three areas of important policy impact: (1) as a tool for understanding and managing cities; (2) as a planning tool for exploring and preparing for potential emergency situations; and (3) as a real-time management tool, drawing on current data as they emerge to create the most reliable picture of the current situation.