ABM get-together May 2015

The May 2015 ABM get-together brought together over twenty Oxford researchers (as well as visitors from Bath and Rutherford Lab) from diverse fields.

Some of the discussion was about how to run large models and experiments using high-performance computers such as those at Oxford’s ARC facility.

Jan Dubbelboer is using NetLogo to create an ABM about flood damage, risk and property buying decisions in London neighbourhoods like Camden. Main challenge is getting the model to run and is about to try running model using HPC here at Oxford (we mentioned Ken’s ARC HPC getting started guide). We also discussed whether FLAME or RePast would be better tools for running models like this with large numbers of agents each of which is running fairly intensive code.

Oliver Lewis from improbable.io gave us an overview of a system that is about to be launched for running very large-scale agent simulations that can be experienced immersively by multiple participants.

Pablo Estevez in Economic Geography is designing a model to study the effect that large oil producers have on the overall economy. In particular whether countries that have large oil production might struggle to diversify their economy i.e. create jobs with a wide range of skills / productive knowledge.

Anders Sandberg of the Future of Humanity Institute gave an update on work he is doing with Feng Zhou on modelling bounded rationality in insurance companies to understand the systemic risk in this industry. This is a “meta-model” since it models the use of models (by insurance companies). Scott Page and the diversity prediction theorem was discussed i.e. whether ideas relating to the wisdom of the crowd and considering a large number of models is relevant.

Matteo Richiardi at the Institute for New Economic Thinking gave us a quick update on recent performance testing he is doing on the new Jas-mine ABM platform.

Justin Lane of Anthropology gave us a last update on his religiosity model before he heads off to Boston to join a very exciting and large international project where he’ll be able to explore these ideas further. Justin is also working out how to move his model to HPC in order to explore the large dimensional space (the model got complicated). Justin is looking at schisms and whether ‘identity’ plays a major role. Adrienne mentioned Michelle Barnes’s work on ethnicity in common-pool resource institutions.

Joanna Bryson from Bath University also managed to make the meeting and gave us an overview of some current exciting research going on within her group. This included work on evolutionary dynamics. She recommended these group and event web sites: http://anagramm.phil.hhu.de/ and http://www.gesis.org/en/events/css-wintersymposium/.

Ken Kahn of IT Services presented slides about NetLogo developments, AgentBase, ABM books, ABM MOOCs, and Rob Axtell’s ABM bibliometrics.

Posted in Meetings | Leave a comment

Leave a Reply