2nd Workshop on Agent-based Modelling at ESCP Europe Berlin
Date
19 May 2017 09:30
- 20 May 2017 18:30
Venue
ESCP Europe Berlin
Download Calendar Event
Agent-based modelling in economics – from toy model to verified tool of analysis.
A platform for discussion + tutorial on the validation and verification of ABM.
The best model of a cat is a cat, preferably the same cat.1 Science needs models to study phenomena which can´t be tested in reality. A 1:1 model would be as complex as the real world and take as long to run. Models thus have to be abstract simplifications. Standard economic mathematical models are high abstractions based on unrealistic behavioral assumptions.
Agent-based models (ABM) have been established as an alternative, offering the potential to integrate empirically observed facts, but economic ABM are still often toy models. They may be helpful to illustrate interrelations or dynamics, but the relevance of their analysis is hard to prove.
The workshop wants to discuss how economic ABM can become more credible by designing them following standard principles of validation and verification. We would like to exchange ideas about methods discussed in literature, such as sensitivity analyses, guidelines for rigor,2 or the “systematic design of experiments” (DOE)3.
We encourage all scientists working with ABM in economics and at least basic experience with model validation to send an abstract of 300-500 words by March 26th 2017 via email to sustbusy@escpeurope.eu, together with the application form available at www.sustbusy.eu. Scientists without a paper may apply explaining their motivation. Active participation is required by agreeing to serve as discussant for a paper.
Notification of acceptance will be given after a review process by March 31, full papers are due by May 5th 2017. The workshop is free of charge.
For more information please click here.
A platform for discussion + tutorial on the validation and verification of ABM.
The best model of a cat is a cat, preferably the same cat.1 Science needs models to study phenomena which can´t be tested in reality. A 1:1 model would be as complex as the real world and take as long to run. Models thus have to be abstract simplifications. Standard economic mathematical models are high abstractions based on unrealistic behavioral assumptions.
Agent-based models (ABM) have been established as an alternative, offering the potential to integrate empirically observed facts, but economic ABM are still often toy models. They may be helpful to illustrate interrelations or dynamics, but the relevance of their analysis is hard to prove.
The workshop wants to discuss how economic ABM can become more credible by designing them following standard principles of validation and verification. We would like to exchange ideas about methods discussed in literature, such as sensitivity analyses, guidelines for rigor,2 or the “systematic design of experiments” (DOE)3.
We encourage all scientists working with ABM in economics and at least basic experience with model validation to send an abstract of 300-500 words by March 26th 2017 via email to sustbusy@escpeurope.eu, together with the application form available at www.sustbusy.eu. Scientists without a paper may apply explaining their motivation. Active participation is required by agreeing to serve as discussant for a paper.
Notification of acceptance will be given after a review process by March 31, full papers are due by May 5th 2017. The workshop is free of charge.
For more information please click here.
Share this story
