Implementation of Mechanism Design and Algorithmic Game theory

#1

Hello,
I have recently studied various topics of Mechanism design and Algorithmic Game theory , thus I really want to use my knowledge of these topics to some implementable stuff.
Since quantecon.py is computational application most of the important fields of economics thus I will be grateful if I can add implementation of mechanism design to it. So it will be helpful if anyone can guide me to start it ?

#2

If you’re interested in game theory, you could also consider looking at our dedicated Games.jl repository. On the Python side, I think this notebook is a good overview of existing support (but things may have changed over the last year.)

In each case, though, the development process would be about the same; build familiarity with the existing codebase, look at relevant issues or natural extensions, check out a fork and add features alongside the appropriate tests and documentation. Then, do a linting pass to make sure the style and such is the same, and submit a PR.

Unfortunately, I don’t think anyone has time to mentor for GSOC or something like that. But if you have specific questions, I’m sure people would be able to answer either here or on the repository itself.

Best of luck.

#3

As far as I know, the people doing algorithmic game theory and that sort of applied mechanism design are mostly in CS and OR departments. I have seen those equilibrium concepts used infrequently among economists (I have seen them in macro occasionally, but for a very different purpose) … But maybe I could be wrong.

Since the whole point of the non-Nash setups is to reduce things linear programming problems, it may make sense to build your software around a serious, commercial solver. Since there are limited complementarities with existing game theory packages, and the dependencies are heavy, my gut is that you are better off creating your own or contributing to one from the CS crew (though I don’t know how to find those).