The investigators involved will be Mark Birkin, Martin Clarke, Alison Heppenstall, Jie Xu, Andy Turner and Paul Towend at Leeds: Mike Batty, Paul Longley, Alan Wilson, Anthony Steed, Andy Hudson-Smith and Richard Milton at UCL.
‘Generative social science’ is widely regarded as one of the grand challenges of the social sciences. The term was popularised by Epstein and Axtell of the Brookings Institution in the book (1996) Growing Artificial Societies: Social Science from the Bottom Up who define it as simulation that “… allows us to grow social structures in silico demonstrating that certain sets of micro-specifications are sufficient to generate the macro-phenomena of interest”. It is consistent with the development of the complexity sciences, with the development of decentralised and distributed agent-based simulation, and with ideas about social and spatial emergence. It requires large-scale data bases for its execution as well as powerful techniques of visualisation for its understanding and dissemination. It provides experimental conditions under which key policy initiatives can be tested on large-scale populations simulated at individual level. It is entirely coincident with the development of e-social science which provides the infrastructure on which such modelling must take place.
The Centre for Advanced Spatial Analysis (CASA) at University College London, and the Centre for Spatial Analysis and Policy (CSAP) at the University of Leeds are collaborating on a proposal to develop this kind of science for the simulation of cities and regions. These two groups have a long history of developing urban and regional models beginning with spatial interaction-allocation models and moving these to aggregate dynamics of urban structure and more recently embracing new approaches at more disaggregate levels which build on microsimulation and agent based modelling. We are proposing to develop a series of demonstrators which will be implemented using e-infrastructures which are emerging from experiments in grid computing and Web 2.0. The interfaces will be highly visual building on CASA's GeoVUE and on CSAP's MoSeS projects which have been funded under ESRC's Digital Social Research.