Workshop: Validation Methods for Agent-Based Models
20 April 2017
The School of Economics will be hosting a workshop on ‘Validation Methods for Agent-Based Models’ on Monday 24 April. The workshop will comprise a series of presentations relating to this very active area of research on the afternoon of 24 April, followed by an informal reception.
Agent-based models (ABM) aim to explore aspects of the interaction of economic agents that are difficult to model using standard approaches, such as the emergence of statistical equilibrium from decentralised interaction. Because these models are typically simulated rather than estimated, a major hurdle facing this approach has traditionally been the problem of calibrating the parameter values of these models from empirical data, known as the 'validation problem'. This problem arises from the emergent properties and non-linearities of ABMs, as well as the fact that each simulation run is idiosyncratic.
There has recently been fast progress on this issue, however, as various methodologies have been developed to address this problem. The purpose of the workshop is to gather these experts in the field of ABM validation in order to examine the relative merits of these methodologies, identify avenues for further progress on the validation problem as well as potential collaborations.
Further details, including the schedule and accompanying papers (where available) can be viewed on our website.
As well as presenting at the workshop, we are pleased that Professor Thomas Lux of the University of Kiel will be staying at the School a little longer as a Thirlwall scholar. Professor Lux's research interests involve the modelling of the dynamics of financial markets, of contagion in interbank and bank-firm networks and of estimation/validation of agent based models. More details can be found via his webpage: https://www.gwif.vwl.uni-kiel.de/en/team/reserach-team/prof.-dr.-thomas-lux
The workshop is sponsored by the Macroeconomics, Growth and History Centre (MaGHiC), University of Kent.