Towards the Design of a Robust Incentive Mechanism for E-Marketplaces with Limited Inventory

In e-marketplaces, reputation systems are prevalently ap-plied to assist buyers to model seller trustworthiness based on ratings shared by other buyers. However, dishonest buy-ers may share untruthful ratings. Many incentive mech-anisms have been proposed to elicit truthful ratings from buyers. These mechanisms have a common implicit assump-tion that sellers can provide a large number of products. In reality, e-markets with limited inventory exist in many scenarios. For example, dentist booking in US, as a mar-ketplace, has been observed the phenomenon that the ser-vice demand is much larger than the service supply, and the second-hand markets where some used and workable goods (e.g., second-hand textbooks) are often in short supply due to lower prices. We call a marketplace in which the demand outweighs the supply a marketplace with limited inventory. In such marketplaces, buyers may lose profit if they pro-vide truthful ratings due to competition from other buyers for the limited products provided by trustworthy sellers. In other words, providing truthful ratings is costly. The ex-isting incentive mechanisms seldom consider such cost im-posed on providing truthful ratings, which motivates us to design a new incentive mechanism for e-marketplaces with limited inventory. Moreover, a general assumption for incen-tive mechanisms is that every agent is rational and has the belief that others are also rational. However, in a realistic scenario, some agents may be irrational (bounded rational-ity or adversary), which may impede the efficiency of the mechanisms. In all, the aim of our research is to design an incentive mechanism for e-marketplaces with limited inven-tory that is robust against irrational agents