My interest for agent-based models in economics goes back to 2008, when I followed a class in computational sociology during my master in Toulouse. Though it was a sociology class, I immediately saw what this methodology could bring to economics – on top of the more traditional equation-based economic theory, not instead.
After this class, I have been advised many times not to engage in agent-based computational economics (ACE) by senior faculties, mostly because it is not an established methodology within economics, and doing so was more or less an academic suicide. But at the end of 2014, after two years in my PhD, I experienced a severe burnout, so severe that I decided to resign. After being convinced that I should not resign and finish it, I had not much to lose by working with a “marginal” methodology. The very notion of risking an academic suicide was not a deterrent anymore. My position was simple: if I should finish my PhD, I should at least enjoy it, and work with a methodology I wanted to explore literally for years1Please, do not derive any advice or lesson from my story. I am just explaining where I come from, I do not intend to influence anyone to do the same choices as the ones I did. Every situation is specific, and should be carefully considered accordingly. The number of factors that influences one’s situation is staggering..
So in 2015, I started to work on ACE. But I was isolated, and it took me a long time before reaching the technological frontier of ACE2I should also mention that as a PhD student in a non-top institution, I self-censored myself a lot. It probably slowed me down.. I received support from established researchers, actually mostly computer scientists, but the hardest part (on top on feeling really isolated) was the lack of comprehensive material to begin with.
The main reference point is Leigh Tesfatsion’s website, which is a great collection of resources. I don’t know why, but I always had a difficult time clicking with her website. Don’t get me wrong: it is a terrific repository, and everyone interested with ACE should have a look at it, and probably start there. But for a reason I cannot figure it out, I was not able to dive in.
I started to collect resources on my own, to answer the (sometimes very specific) questions I had. But why should I keep all this material just for me, whereas other (junior, and not so junior) researchers could also benefit from accessing it?
This is why I am announcing a wiki entirely dedicated to ACE. Because it involves a growing field of study, this wiki is intended as a never-ending project. I will build it with beginners in mind, especially economists – but I think that non-economists could also derive some useful tips from reading it.
I do not intend this wiki as a competitor to Leigh’s comprehensive material, but as a complement. I see it as another window pointing at the same material and ideas. And as it is a scientific project, I will obviously quote the relevant authors and works, including Leigh’s website.
You can find this wiki within the general wiki on my website.
So far, I already published a page on the ODD+D protocol, which is a standardized way to describe agent-based models. If you want to have an idea of the kind of resources I will gather, and how I will organize them, you can have a look here.
Feel free to comment and/or offer suggestions! I will reference the author of any suggestion I will implement on the website. Even though hosted on my website, this wiki is not intended to be Olivier Simard-Casanova-centric. The goal really is to facilitate even more the access to ACE.