Characterizing Truthfulness In Discrete Domains

Algorithmic mechanism design [9; 10] focuses on the design of algorithms that aim to achieve global objectives in settings in which the "input" is provided by self-interested strategic players2. This necessitates the design of algorithms that are incentive-compatible (a.k.a. truthful 3) in the sense that players are incentivized via payments to behave as instructed. The most natural approach to designing incentive-compatible algorithms is coming up with an algorithm and an explicit payment scheme that guarantees its incentive-compatibility. However, finding appropriate payments is often a difficult, setting-specific, task, which is mostly achievable for very simple types of algorithms.