Algorithmic mechanism design
Designing markets and games to achieve what we collectively want from what we individually want
September 22, 2014 — February 27, 2024
Hard theory of incentive mechanisms, where we can plug numbers into sufficiently abstract models and maybe extract computational complexity results, and get economic optimality as a side effect. Or the other way around.
Every blockchain-style cryptowhatsit is a mechanism design problem. Better governance is a mechanism design problem.
1 Fair division
Especially cake-cutting.
spliddit, a website to use optimal cake cutting algorithms to allocate credit/rent/whatever (Goldman and Procaccia 2015).Oops their site went down (N. Shah 2017). Code here: nisarg89/spliddit
Alternative: the New York Times rent calculator, as mentioned in Albert Sun’s article Sun (2014) about Su’s research into Sperner’s lemma (Su 1999).
Erica Klarreich’s exposé: The Mathematics of Cake Cutting
2 Tutorials
Aaron Roth’s Algorithmic Game theory course
In this course, we will take an algorithmic perspective on problems in game theory. We will consider questions such as: how should an auction for scarce goods be structured if the seller wishes to maximize his revenue? How badly will traffic be snarled if drivers each selfishly try to minimize their commute time, compared to if a benevolent dictator directed traffic? How can couples be paired so that no two couples wish to swap partners in hindsight? How can you be as successful at betting on horse races as the best horse racing expert, without knowing anything about horse racing? How can we set prices so that all goods get sold, and everyone gets their favorite good?
3 Incoming
player vs game
Mechanism Design on the AI Alignment Forum
Related Pages: Game Theory, Incentives, Principal-Agent Problems, Cryptocurrencies and blockchain, Public discourse
Related Sequences: Mechanism Design