Tuesday, February 13, 2007

Improving Evidence Based Policy Decisions

Improving Evidence Based Policy Decisions: Piloting the Application of Advanced Computer Modelling Techniques to Real Life Policy Problems

The other day we attended a workshop organised by Georgios Theodoropoulos from the University of Birmingham called Improving Evidence Based Policy Decisions

We gave a talk entitled ‘Validating and Verifying Agent-Based Models for Planning and Public Policy Analysis’ which focused around validation and verification of ABM (issues which will be explored in later posts). The presentation highlighted how urban modelling has changed, especially how we conceptualise systems; a move from aggregate, to disaggregate; focusing on change not equilibrium, where dynamics has come to the fore. Thus a crucial issue is how do we know how good this new class of models is in terms of their fit to reality? Do we even need to worry about this any longer? These issues where explored using ABM developed at CASA from fine scale pedestrian models to more aggregate models examining residential segregation and urban sprawl.

Other talks examined the role data driven simulation for housing policy which is part of the Adaptive Intelligent Model-Building for the Social Sciences using Symbiotic Simulation (AIMSS) project. While Mark Birkin gave a talk entitled ‘Issues in the specification and validation of a complex socio-demographic simulation model’ which showed the MoSeS (Modelling and Simulation for e-Social Science) project from the University of Leeds which aims to develop a demographic simulation at the level of individuals and households to give robust forecasts of the future population of the UK based on microsimulation techniques but in a dynamic context.

A copy of the presentation Mike Batty and I gave can be downloaded by clicking here. Any comments more than welcome.
For details of other talks (including pdfs) see the National Centre for e-Social Science (NCeSS) website

No comments: