Mechanism design is a branch of economics focused on creating incentives to encourage the adoption of behavior such that higher-level goals are achieved. Mechanism designers begin by applying reverse game theory to construct mathematical descriptions of incentives, actors, and game mechanics. How can a proposed mechanism be validated? In a simulation, agents modeled on heuristics or control theory can interact with a mechanism and provide first-order checks. Further, given sufficient compute resources, reinforcement learning can be used to obtain performant or adversarial agents that provide robust experimental validation.

This workshop will consist of short talks and tutorials focused on the roles of and interplay between optimization, control theory, and reinforcement learning in the mechanism design life cycle. Presenters are encouraged to share how they have used optimization, control theory, reinforcement learning, and related techniques to build or analyze real-world cryptoeconomic systems. Regardless of prior background, we want attendees to leave with intuition regarding which tools and approaches work well and in what situations. We also seek to identify open problems in Ethereum where the above-mentioned techniques can be applied.