skip to primary navigationskip to content

Alison Heppenstall and Nicholas Malleson: Abstract and background reading

Title image Alison Heppenstall

SHORT ABSTACT: The past decade has seen a rise in the popularity of agent-based models for simulating how individuals navigate and experience cities.  This has been in part, powered through innovations in new forms of data and the emergence of machine-learning inspired algorithms.  We are now approaching the possibility of identifying key processes driving cities and creating realistic ‘digital twins’.  One of the most exciting opportunities offered by these innovations is the potential for creating simulations that can assimilate real-time data to produce highly accurate short-term forecasts.  This presentation will showcase advances (and the issues) with creating dynamic agent-based simulations of cities. 

READ MORE: 

The past decade has seen a rise in the popularity of agent-based models for simulating how individuals navigate and experience cities.  This has been driven by a number of factors including firstly, changes about how we conceptualise cities.  Both Batty (2013) and O’Sullivan (2012) describe cities as being the product of networks (transmitting information, money, ideas) with decisions made by individuals driving many of the processes.  Secondly, viewing the city through an individual perspective has led to an upsurge in the development and application of individual-based methods, most prominently, agent-based modelling.  Finally, the appearance of new forms of ‘big data’ has enabled these models to operate more ‘realistic’ rules and run at finer levels of granularity.  

We are now approaching the possibility of identifying key processes driving cities and creating realistic ‘digital twins’.  One of the most exciting opportunities offered by these innovations is the potential for creating simulations that can assimilate real-time data to produce highly accurate short-term forecasts.  Whilst assimilation approaches are established in other disciplines such as metereology, their use in the social sciences is still limited. These approaches could be extremely valuable for both understanding how individuals react in real-time to emergency situations and helping to quantify the uncertainty with our simulation outputs.  

This presentation will showcase advances (and the issues) with creating dynamic agent-based simulations of cities. 

BACKGROUND READING

Heppenstall A, Malleson N. 2020. Building cities from slime mould, agents and quantum field theory.  19th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2020) Proc. of the 19th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2020

Malleson, N., K. Minors, Le-Minh Kieu , J. A. Ward , A. West and A. Heppenstall (2020) Simulating Crowds in Real Time with Agent-Based Modelling and a Particle Filter. Journal of Artificial Societies and Social Simulation (JASSS) 23 (3). http://jasss.soc.surrey.ac.uk/23/3/3.html DOI: 10.18564/jasss.4266 (open access) 

Manson S, An L, Clarke KC, Heppenstall A, Koch J, Krzyzanowski B, Morgan F, O'Sullivan D, Runck BC, Shook E, Tesfatsion L. 2020. Methodological Issues of Spatial Agent-Based Models.  Journal of Artificial Societies and Social Simulation.  23(1) 

Kieu, Le-Minh, N. Malleson, and A. Heppenstall (2019). Dealing with Uncertainty in Agent-Based Models for Short-Term Predictions’. Royal Society Open Science 7(1): 191074. DOI: 10.1098/rsos.191074 (open access) 

Crooks, A, N. Malleson, E. Manley, A. Heppenstall (2019) Agent-Based Modelling and Geographical Information Systems. Sage.