|Location:||SLMath: Eisenbud Auditorium, Online/Virtual|
Contract Design Under Uncertainty
As algorithms increasingly interact with rational, self-interested individuals, their design must take participants’ incentives into account. The traditional focus in algorithmic game theory literature has been the design of algorithms that incentivize truthfulness (as seen, for example, in ad auctions). In this talk, we shift focus to algorithms that incentivize effort, aligning with the economic discipline of contract design.
The goal of this talk is to address some generalizations of the classic contract design model and demonstrate how algorithmic approaches contribute to this research domain. Our primary focus will be a natural contractual problem involving both aspects of effort and truthfulness incentivization. We will see that the optimal mechanisms can be complex and unintuitive, and do not resemble contracts used in practice. We then shift perspective to simple contracts, such as linear or commission-based contracts, and demonstrate that they are close to optimal under certain natural conditions. This provides a justification for the prevalence of simple contracts in practical applications.