Large-Scale Modeling of Economic Systems
aMike Holcombe, Simon Coakley, Mariam Kiran
bShawn Chin, Chris Greenough, David Worth
cSilvano Cincotti, M. Raberto
dAndrea Teglio
eChristophe Deissenberg
fSander van der Hoog, Herbert Dawid, Simon Gemkow, Philipp Harting
gMichael Neugart
aDepartment of Computer Science, University of Sheffield
bSoftware Engineering Group, Rutherford Appleton Laboratory, STFC
cDIME-CINEF, University of Genoa
dDepartment of Economics, Universitat Jaume I
eFaculty of Economics, Université de la Méditerranée (Aix-Marseille II)
fDepartment of Economics, Bielefeld University
gFachgebiet Finanzwissenschaft und Wirtschaftspolitik, Technische Universität Darmstadt
Abstract
Following the events of the credit crunch and the onset of a global recession, alternative ways of modeling modern economies and mechanisms for carrying out policy analysis are now an urgent priority. Traditional mathematical economics is widely viewed to have been compromised through gross simplifications with many assumptions that are now seen to be unjustified. New ways of looking at economics that are more grounded in reality are required, and agent-based computational economics is now receiving a lot of attention. Although the ideas are not new, the previous attempts to use this approach have been largely limited by the inability to model realistically large systems with millions of complex agents. Without this capability, the usefulness of the approach is limited. The EU-STREP project EURACE brought together a consortium of leading economists, experts in parallel supercomputing, and the designers of the FLAME framework to build the largest and most complete model of the European Union economy ever built.
